Builtin Functions Reference

Table of Contents

Introduction

The DML (Declarative Machine Learning) language has built-in functions which enable access to both low- and high-level functions to support all kinds of use cases.

A builtin is either implemented on a compiler level or as DML scripts that are loaded at compile time.

Built-In Construction Functions

There are some functions which generate an object for us. They create matrices, tensors, lists and other non-primitive objects.

tensor-Function

The tensor-function creates a tensor for us.

tensor(data, dims, byRow = TRUE)

Arguments

Name Type Default Description
data Matrix[?], Tensor[?], Scalar[?] required The data with which the tensor should be filled. See data-Argument.
dims Matrix[Integer], Tensor[Integer], Scalar[String], List[Integer] required The dimensions of the tensor. See dims-Argument.
byRow Boolean TRUE NOT USED. Will probably be removed or replaced.

Note that this function is highly unstable and will be overworked and might change signature and functionality.

Returns

Type Description
Tensor[?] The generated Tensor. Will support more datatypes than Double.
data-Argument

The data-argument can be a Matrix of any datatype from which the elements will be taken and placed in the tensor until filled. If given as a Tensor the same procedure takes place. We iterate through Matrix and Tensor by starting with each dimension index at 0 and then incrementing the lowest one, until we made a complete pass over the dimension, and then increasing the dimension index above. This will be done until the Tensor is completely filled.

If data is a Scalar, we fill the whole tensor with the value.

dims-Argument

The dimension of the tensor can either be given by a vector represented by either by a Matrix, Tensor, String or List. Dimensions given by a String will be expected to be concatenated by spaces.

Example

print("Dimension matrix:");
d = matrix("2 3 4", 1, 3);
print(toString(d, decimal=1))

print("Tensor A: Fillvalue=3, dims=2 3 4");
A = tensor(3, d); # fill with value, dimensions given by matrix
print(toString(A))

print("Tensor B: Reshape A, dims=4 2 3");
B = tensor(A, "4 2 3"); # reshape tensor, dimensions given by string
print(toString(B))

print("Tensor C: Reshape dimension matrix, dims=1 3");
C = tensor(d, list(1, 3)); # values given by matrix, dimensions given by list
print(toString(C, decimal=1))

print("Tensor D: Values=tst, dims=Tensor C");
D = tensor("tst", C); # fill with string, dimensions given by tensor
print(toString(D))

Note that reshape construction is not yet supported for SPARK execution.

DML-Bodied Built-In Functions

DML-bodied built-in functions are written as DML-Scripts and executed as such when called.

confusionMatrix-Function

A confusionMatrix-accepts a vector for prediction and a one-hot-encoded matrix, then it computes the max value of each vector and compare them, after which it calculates and returns the sum of classifications and the average of each true class.

Usage

confusionMatrix(P, Y)

Arguments

Name Type Default Description
P Matrix[Double] vector of prediction
Y Matrix[Double] vector of Golden standard One Hot Encoded

Returns

Name Type Description
ConfusionSum Matrix[Double] The Confusion Matrix Sums of classifications
ConfusionAvg Matrix[Double] The Confusion Matrix averages of each true class

Example

numClasses = 1
z = rand(rows = 5, cols = 1, min = 1, max = 9)
X = round(rand(rows = 5, cols = 1, min = 1, max = numClasses))
y = toOneHot(X, numClasses)
[ConfusionSum, ConfusionAvg] = confusionMatrix(P=z, Y=y)

correctTypos-Function

The correctTypos - function tries to correct typos in a given frame. This algorithm operates on the assumption that most strings are correct and simply swaps strings that do not occur often with similar strings that occur more often. If correct is set to FALSE only prints suggested corrections without affecting the frame.

Usage

correctTypos(strings, frequency_threshold, distance_threshold, decapitalize, correct, is_verbose)

Arguments

NAME TYPE DEFAULT Description
strings String The nx1 input frame of corrupted strings
frequency_threshold Double 0.05 Strings that occur above this relative frequency level will not be corrected
distance_threshold Int 2 Max editing distance at which strings are considered similar
decapitalize Boolean TRUE Decapitalize all strings before correction
correct Boolean TRUE Correct strings or only report potential errors
is_verbose Boolean FALSE Print debug information

Returns

TYPE Description
String Corrected nx1 output frame

Example

A = read(β€œfile1”, data_type=”frame”, rows=2000, cols=1, format=”binary”)
A_corrected = correctTypos(A, 0.02, 3, FALSE, TRUE)

cspline-Function

This cspline-function solves Cubic spline interpolation. The function usages natural spline with \(q_1''(x_0) == q_n''(x_n) == 0.0\). By default, it calculates via csplineDS-function.

Algorithm reference: https://en.wikipedia.org/wiki/Spline_interpolation#Algorithm_to_find_the_interpolating_cubic_spline

Usage

[result, K] = cspline(X, Y, inp_x, tol, maxi)

Arguments

Name Type Default Description
X Matrix[Double] 1-column matrix of x values knots. It is assumed that x values are monotonically increasing and there is no duplicate points in X
Y Matrix[Double] 1-column matrix of corresponding y values knots
inp_x Double the given input x, for which the cspline will find predicted y
mode String DS Specifies that method for cspline (DS - Direct Solve, CG - Conjugate Gradient)
tol Double -1.0 Tolerance (epsilon); conjugate gradient procedure terminates early if L2 norm of the beta-residual is less than tolerance * its initial norm
maxi Integer -1 Maximum number of conjugate gradient iterations, 0 = no maximum

Returns

Name Type Description
pred_Y Matrix[Double] Predicted values
K Matrix[Double] Matrix of k parameters

Example

num_rec = 100 # Num of records
X = matrix(seq(1,num_rec), num_rec, 1)
Y = round(rand(rows = 100, cols = 1, min = 1, max = 5))
inp_x = 4.5
tolerance = 0.000001
max_iter = num_rec
[result, K] = cspline(X=X, Y=Y, inp_x=inp_x, tol=tolerance, maxi=max_iter)

csplineCG-Function

This csplineCG-function solves Cubic spline interpolation with conjugate gradient method. Usage will be same as cspline-function.

Usage

[result, K] = csplineCG(X, Y, inp_x, tol, maxi)

Arguments

Name Type Default Description
X Matrix[Double] 1-column matrix of x values knots. It is assumed that x values are monotonically increasing and there is no duplicate points in X
Y Matrix[Double] 1-column matrix of corresponding y values knots
inp_x Double the given input x, for which the cspline will find predicted y
tol Double -1.0 Tolerance (epsilon); conjugate gradient procedure terminates early if L2 norm of the beta-residual is less than tolerance * its initial norm
maxi Integer -1 Maximum number of conjugate gradient iterations, 0 = no maximum

Returns

Name Type Description
pred_Y Matrix[Double] Predicted values
K Matrix[Double] Matrix of k parameters

Example

num_rec = 100 # Num of records
X = matrix(seq(1,num_rec), num_rec, 1)
Y = round(rand(rows = 100, cols = 1, min = 1, max = 5))
inp_x = 4.5
tolerance = 0.000001
max_iter = num_rec
[result, K] = csplineCG(X=X, Y=Y, inp_x=inp_x, tol=tolerance, maxi=max_iter)

csplineDS-Function

This csplineDS-function solves Cubic spline interpolation with direct solver method.

Usage

[result, K] = csplineDS(X, Y, inp_x)

Arguments

Name Type Default Description
X Matrix[Double] 1-column matrix of x values knots. It is assumed that x values are monotonically increasing and there is no duplicate points in X
Y Matrix[Double] 1-column matrix of corresponding y values knots
inp_x Double the given input x, for which the cspline will find predicted y

Returns

Name Type Description
pred_Y Matrix[Double] Predicted values
K Matrix[Double] Matrix of k parameters

Example

num_rec = 100 # Num of records
X = matrix(seq(1,num_rec), num_rec, 1)
Y = round(rand(rows = 100, cols = 1, min = 1, max = 5))
inp_x = 4.5

[result, K] = csplineDS(X=X, Y=Y, inp_x=inp_x)

cvlm-Function

The cvlm-function is used for cross-validation of the provided data model. This function follows a non-exhaustive cross validation method. It uses lm and lmPredict functions to solve the linear regression and to predict the class of a feature vector with no intercept, shifting, and rescaling.

Usage

cvlm(X, y, k)

Arguments

Name Type Default Description
X Matrix[Double] required Recorded Data set into matrix
y Matrix[Double] required 1-column matrix of response values.
k Integer required Number of subsets needed, It should always be more than 1 and less than nrow(X)
icpt Integer 0 Intercept presence, shifting and rescaling the columns of X
reg Double 1e-7 Regularization constant (lambda) for L2-regularization. set to nonzero for highly dependant/sparse/numerous features

Returns

Type Description
Matrix[Double] Response values
Matrix[Double] Validated data set

Example

X = rand (rows = 5, cols = 5)
y = X %*% rand(rows = ncol(X), cols = 1)
[predict, beta] = cvlm(X = X, y = y, k = 4)

DBSCAN-Function

The dbscan() implements the DBSCAN Clustering algorithm using Euclidian distance.

Usage

Y = dbscan(X = X, eps = 2.5, minPts = 5)

Arguments

Name Type Default Description
X Matrix[Double] required The input Matrix to do DBSCAN on.
eps Double 0.5 Maximum distance between two points for one to be considered reachable for the other.
minPts Int 5 Number of points in a neighborhood for a point to be considered as a core point (includes the point itself).

Returns

Type Description
Matrix[Integer] The mapping of records to clusters
Matrix[Double] The coordinates of all points considered part of a cluster

Example

X = rand(rows=1780, cols=180, min=1, max=20) 
[indices, model] = dbscan(X = X, eps = 2.5, minPts = 360)

decisionTree-Function

The decisionTree() implements the classification tree with both scale and categorical features.

Usage

M = decisionTree(X, Y, R);

Arguments

Name Type Default Description
X Matrix[Double] required Feature matrix X; note that X needs to be both recoded and dummy coded
Y Matrix[Double] required Label matrix Y; note that Y needs to be both recoded and dummy coded
R Matrix[Double] ” “ Matrix R which for each feature in X contains the following information
- R[1,]: Row Vector which indicates if feature vector is scalar or categorical. 1 indicates
a scalar feature vector, other positive Integers indicate the number of categories
If R is not provided by default all variables are assumed to be scale
bins Integer 20 Number of equiheight bins per scale feature to choose thresholds
depth Integer 25 Maximum depth of the learned tree
verbose Boolean FALSE boolean specifying if the algorithm should print information while executing

Returns

Name Type Description
M Matrix[Double] Each column of the matrix corresponds to a node in the learned tree

Example

X = matrix("4.5 4.0 3.0 2.8 3.5
            1.9 2.4 1.0 3.4 2.9
            2.0 1.1 1.0 4.9 3.4
            2.3 5.0 2.0 1.4 1.8
            2.1 1.1 3.0 1.0 1.9", rows=5, cols=5)
Y = matrix("1.0
            0.0
            0.0
            1.0
            0.0", rows=5, cols=1)
R = matrix("1.0 1.0 3.0 1.0 1.0", rows=1, cols=5)
M = decisionTree(X = X, Y = Y, R = R)

discoverFD-Function

The discoverFD-function finds the functional dependencies.

Usage

discoverFD(X, Mask, threshold)

Arguments

Name Type Default Description
X Double Input Matrix X, encoded Matrix if data is categorical
Mask Double A row vector for interested features i.e. Mask =[1, 0, 1] will exclude the second column from processing
threshold Double threshold value in interval [0, 1] for robust FDs

Returns

Type Description
Double matrix of functional dependencies

dist-Function

The dist-function is used to compute Euclidian distances between N d-dimensional points.

Usage

dist(X)

Arguments

Name Type Default Description
X Matrix[Double] required (n x d) matrix of d-dimensional points

Returns

Type Description
Matrix[Double] (n x n) symmetric matrix of Euclidian distances

Example

X = rand (rows = 5, cols = 5)
Y = dist(X)

dmv-Function

The dmv-function is used to find disguised missing values utilising syntactical pattern recognition.

Usage

dmv(X, threshold, replace)

Arguments

Name Type Default Description
X Frame[String] required Input Frame
threshold Double 0.8 threshold value in interval [0, 1] for dominant pattern per column (e.g., 0.8 means that 80% of the entries per column must adhere this pattern to be dominant)
replace String “NA” The string disguised missing values are replaced with

Returns

Type Description
Frame[String] Frame X including detected disguised missing values

Example

A = read("fileA", data_type="frame", rows=10, cols=8);
Z = dmv(X=A)
Z = dmv(X=A, threshold=0.9)
Z = dmv(X=A, threshold=0.9, replace="NaN")

ffPredict-Function

The ffPredict-function computes prediction of the given thata using trained model.

It takes the list of layers returned by the ffTrain-function.

Usage

prediction = ffPredict(model, x_test, 128)

Arguments

Name Type Default Description
Model List[unknown] required ffTrain model
X Matrix[Double] required Data for making prediction
batch_size Integer 128 Batch Size

Returns

Name Type Description
pred Matrix[Double] Predictions

Example

model = ffTrain(x_train, y_train, 128, 10, 0.001, ...)

prediction = ffPredict(model=model, X=x_test)

ffTrain-Function

The ffTrain-function trains simple feed-forward neural network.

Neural network trained has the following architecture:

affine1 -> relu -> dropout -> affine2 -> configurable output layer activation

Hidden layer has 128 neurons. Dropout rate is 0.35. Input and ouptut sizes are inferred from X and Y.

Usage

model = ffTrain(X=x_train, Y=y_train, out_activation="sigmoid", loss_fcn="cel")

Arguments

Name Type Default Description
X Matrix[Double] required Training data
Y Matrix[Double] required Labels/Target values
batch_size Integer 64 Batch size
epochs Integer 20 Number of epochs
learning_rate Double 0.003 Learning rate
out_activation String required User specified ouptut layer activation
loss_fcn String required User specified loss function
shuffle Boolean False Shuffle dataset
validation_split Double 0.0 Fraction of dataset to be used as validation set
seed Integer -1 Seed for model initialization. Seed value of -1 will generate random seed
verbose Boolean False Flag indicates if function should print to stdout

User can specify following output layer activations: sigmoid, relu, lrelu, tanh, softmax, logits (no activation).

User can specify following loss functions: l1, l2, log_loss, logcosh_loss, cel (cross-entropy loss).

When validation set is used function outputs validation loss to the stdout after each epoch.

Returns

Name Type Description
model List[unknown] Trained model containing weights of affine layers and activation of output layer

Example

model = ffTrain(X=x_train, Y=y_train, batch_size=128, epochs=10, learning_rate=0.001, out_activation="sigmoid", loss_fcn="cel", shuffle=TRUE, validation_split=0.2, verbose=TRUE)

prediction = ffPredict(model=model, X=x_train)

gaussianClassifier-Function

The gaussianClassifier-function computes prior probabilities, means, determinants, and inverse covariance matrix per class.

Classification is as per \(p(C=c | x) = p(x | c) * p(c)\) Where \(p(x | c)\) is the (multivariate) Gaussian P.D.F. for class \(c\), and \(p(c)\) is the prior probability for class \(c\).

Usage

[prior, means, covs, det] = gaussianClassifier(D, C, varSmoothing)

Arguments

Name Type Default Description
D Matrix[Double] required Input matrix (training set
C Matrix[Double] required Target vector
varSmoothing Double 1e-9 Smoothing factor for variances
verbose Boolean TRUE Print accuracy of the training set

Returns

Name Type Description
classPriors Matrix[Double] Vector storing the class prior probabilities
classMeans Matrix[Double] Matrix storing the means of the classes
classInvCovariances List[Unknown] List of inverse covariance matrices
determinants Matrix[Double] Vector storing the determinants of the classes

Example


X = rand (rows = 200, cols = 50 )
y = X %*% rand(rows = ncol(X), cols = 1)

[prior, means, covs, det] = gaussianClassifier(D=X, C=y, varSmoothing=1e-9)

glm-Function

The glm-function is a flexible generalization of ordinary linear regression that allows for response variables that have error distribution models.

Usage

glm(X,Y)

Arguments

Name Type Default Description
X Matrix[Double] required matrix X of feature vectors
Y Matrix[Double] required matrix Y with either 1 or 2 columns: if dfam = 2, Y is 1-column Bernoulli or 2-column Binomial (#pos, #neg)
dfam Int 1 Distribution family code: 1 = Power, 2 = Binomial
vpow Double 0.0 Power for Variance defined as (mean)^power (ignored if dfam != 1): 0.0 = Gaussian, 1.0 = Poisson, 2.0 = Gamma, 3.0 = Inverse Gaussian
link Int 0 Link function code: 0 = canonical (depends on distribution), 1 = Power, 2 = Logit, 3 = Probit, 4 = Cloglog, 5 = Cauchit
lpow Double 1.0 Power for Link function defined as (mean)^power (ignored if link != 1): -2.0 = 1/mu^2, -1.0 = reciprocal, 0.0 = log, 0.5 = sqrt, 1.0 = identity
yneg Double 0.0 Response value for Bernoulli “No” label, usually 0.0 or -1.0
icpt Int 0 Intercept presence, X columns shifting and rescaling: 0 = no intercept, no shifting, no rescaling; 1 = add intercept, but neither shift nor rescale X; 2 = add intercept, shift & rescale X columns to mean = 0, variance = 1
reg Double 0.0 Regularization parameter (lambda) for L2 regularization
tol Double 1e-6 Tolerance (epislon) value.
disp Double 0.0 (Over-)dispersion value, or 0.0 to estimate it from data
moi Int 200 Maximum number of outer (Newton / Fisher Scoring) iterations
mii Int 0 Maximum number of inner (Conjugate Gradient) iterations, 0 = no maximum

Returns

Type Description
Matrix[Double] Matrix whose size depends on icpt ( icpt=0: ncol(X) x 1; icpt=1: (ncol(X) + 1) x 1; icpt=2: (ncol(X) + 1) x 2)

Example

X = rand (rows = 5, cols = 5 )
y = X %*% rand(rows = ncol(X), cols = 1)
beta = glm(X=X,Y=y)

gmm-Function

The gmm-function implements builtin Gaussian Mixture Model with four different types of covariance matrices i.e., VVV, EEE, VVI, VII and two initialization methods namely “kmeans” and “random”.

Usage

gmm(X=X, n_components = 3,  model = "VVV",  init_params = "random", iter = 100, reg_covar = 0.000001, tol = 0.0001, verbose=TRUE)

Arguments

Name Type Default Description
X Double Matrix X of feature vectors.
n_components Integer 3 Number of n_components in the Gaussian mixture model
model String “VVV” “VVV”: unequal variance (full),each component has its own general covariance matrix

“EEE”: equal variance (tied), all components share the same general covariance matrix

“VVI”: spherical, unequal volume (diag), each component has its own diagonal covariance matrix

“VII”: spherical, equal volume (spherical), each component has its own single variance
init_param String “kmeans” initialize weights with “kmeans” or “random”
iterations Integer 100 Number of iterations
reg_covar Double 1e-6 regularization parameter for covariance matrix
tol Double 0.000001 tolerance value for convergence
verbose Boolean False Set to true to print intermediate results.

Returns

Name Type Default Description
weight Double A matrix whose [i,k]th entry is the probability that observation i in the test data belongs to the kth class
labels Double Prediction matrix
df Integer Number of estimated parameters
bic Double Bayesian information criterion for best iteration

Example

X = read($1)
[labels, df, bic] = gmm(X=X, n_components = 3,  model = "VVV",  init_params = "random", iter = 100, reg_covar = 0.000001, tol = 0.0001, verbose=TRUE)

gnmf-Function

The gnmf-function does Gaussian Non-Negative Matrix Factorization. In this, a matrix X is factorized into two matrices W and H, such that all three matrices have no negative elements. This non-negativity makes the resulting matrices easier to inspect.

Usage

gnmf(X, rnk, eps = 10^-8, maxi = 10)

Arguments

Name Type Default Description
X Matrix[Double] required Matrix of feature vectors.
rnk Integer required Number of components into which matrix X is to be factored.
eps Double 10^-8 Tolerance
maxi Integer 10 Maximum number of conjugate gradient iterations.

Returns

Type Description
Matrix[Double] List of pattern matrices, one for each repetition.
Matrix[Double] List of amplitude matrices, one for each repetition.

Example

X = rand(rows = 50, cols = 10)
W = rand(rows = nrow(X), cols = 2, min = -0.05, max = 0.05);
H = rand(rows = 2, cols = ncol(X), min = -0.05, max = 0.05);
gnmf(X = X, rnk = 2, eps = 10^-8, maxi = 10)

gridSearch-Function

The gridSearch-function is used to find the optimal hyper-parameters of a model which results in the most accurate predictions. This function takes train and eval functions by name.

Usage

gridSearch(X, y, train, predict, params, paramValues, verbose)

Arguments

Name Type Default Description
X Matrix[Double] required Input Matrix of vectors.
y Matrix[Double] required Input Matrix of vectors.
train String required Specified training function.
predict String required Evaluation based function.
params List[String] required List of parameters
paramValues List[Unknown] required Range of values for the parameters
verbose Boolean TRUE If TRUE print messages are activated

Returns

Type Description
Matrix[Double] Parameter combination
Frame[Unknown] Best results model

Example

X = rand (rows = 50, cols = 10)
y = X %*% rand(rows = ncol(X), cols = 1)
params = list("reg", "tol", "maxi")
paramRanges = list(10^seq(0,-4), 10^seq(-5,-9), 10^seq(1,3))
[B, opt]= gridSearch(X=X, y=y, train="lm", predict="lmPredict", params=params, paramValues=paramRanges, verbose = TRUE)

hyperband-Function

The hyperband-function is used for hyper parameter optimization and is based on multi-armed bandits and early elimination. Through multiple parallel brackets and consecutive trials it will return the hyper parameter combination which performed best on a validation dataset. A set of hyper parameter combinations is drawn from uniform distributions with given ranges; Those make up the candidates for hyperband. Notes:

Usage

hyperband(X_train, y_train, X_val, y_val, params, paramRanges, R, eta, verbose)

Arguments

Name Type Default Description
X_train Matrix[Double] required Input Matrix of training vectors.
y_train Matrix[Double] required Labels for training vectors.
X_val Matrix[Double] required Input Matrix of validation vectors.
y_val Matrix[Double] required Labels for validation vectors.
params List[String] required List of parameters to optimize.
paramRanges Matrix[Double] required The min and max values for the uniform distributions to draw from. One row per hyper parameter, first column specifies min, second column max value.
R Scalar[int] 81 Controls number of candidates evaluated.
eta Scalar[int] 3 Determines fraction of candidates to keep after each trial.
verbose Boolean TRUE If TRUE print messages are activated.

Returns

Type Description
Matrix[Double] 1-column matrix of weights of best performing candidate
Frame[Unknown] hyper parameters of best performing candidate

Example

X_train = rand(rows=50, cols=10);
y_train = rowSums(X_train) + rand(rows=50, cols=1);
X_val = rand(rows=50, cols=10);
y_val = rowSums(X_val) + rand(rows=50, cols=1);

params = list("reg");
paramRanges = matrix("0 20", rows=1, cols=2);

[bestWeights, optHyperParams] = hyperband(X_train=X_train, y_train=y_train, 
    X_val=X_val, y_val=y_val, params=params, paramRanges=paramRanges);

img_brightness-Function

The img_brightness-function is an image data augumentation function. It changes the brightness of the image.

Usage

img_brightness(img_in, value, channel_max)

Arguments

Name Type Default Description
img_in Matrix[Double] Input matrix/image
value Double The amount of brightness to be changed for the image
channel_max Integer Maximum value of the brightness of the image

Returns

Name Type Default Description
img_out Matrix[Double] Output matrix/image

Example

A = rand(rows = 3, cols = 3, min = 0, max = 255)
B = img_brightness(img_in = A, value = 128, channel_max = 255)

img_crop-Function

The img_crop-function is an image data augumentation function. It cuts out a subregion of an image.

Usage

img_crop(img_in, w, h, x_offset, y_offset)

Arguments

Name Type Default Description
img_in Matrix[Double] Input matrix/image
w Integer The width of the subregion required
h Integer The height of the subregion required
x_offset Integer The horizontal coordinate in the image to begin the crop operation
y_offset Integer The vertical coordinate in the image to begin the crop operation

Returns

Name Type Default Description
img_out Matrix[Double] Cropped matrix/image

Example

A = rand(rows = 3, cols = 3, min = 0, max = 255) 
B = img_crop(img_in = A, w = 20, h = 10, x_offset = 0, y_offset = 0)

img_mirror-Function

The img_mirror-function is an image data augumentation function. It flips an image on the X (horizontal) or Y (vertical) axis.

Usage

img_mirror(img_in, horizontal_axis)

Arguments

Name Type Default Description
img_in Matrix[Double] Input matrix/image
horizontal_axis Boolean If TRUE, the image is flipped with respect to horizontal axis otherwise vertical axis

Returns

Name Type Default Description
img_out Matrix[Double] Flipped matrix/image

Example

A = rand(rows = 3, cols = 3, min = 0, max = 255)
B = img_mirror(img_in = A, horizontal_axis = TRUE)

impurityMeasures-Function

impurityMeasures() computes the measure of impurity for each feature of the given dataset based on the passed method (gini or entropy).

Usage

IM = impurityMeasures(X = X, Y = Y, R = R, n_bins = 20, method = "gini");

Arguments

Name Type Default Description
X Matrix[Double] Feature matrix X
Y Matrix[Double] Target vector Y containing only 0 or 1 values
R Matrix[Double] Row vector R indicating whether a feature is categorical or continuous. 1 denotes a continuous feature, 2 denotes a categorical feature.
n_bins Integer 20 Number of equi-width bins for binning in case of scale features.
method String String indicating the method to use; either “entropy” or “gini”.

Returns

Name Type Description
IM Matrix[Double] (1 x ncol(X)) row vector containing information/gini gain for each feature of the dataset. In case of gini, the values denote the gini gains, i.e. how much impurity was removed with the respective split. The higher the value, the better the split. In case of entropy, the values denote the information gain, i.e. how much entropy was removed. The higher the information gain, the better the split.

Example

X = matrix("4.0 3.0 2.8 3.5
            2.4 1.0 3.4 2.9
            1.1 1.0 4.9 3.4
            5.0 2.0 1.4 1.8
            1.1 3.0 1.0 1.9", rows=5, cols=4)
Y = matrix("1.0
            0.0
            0.0
            1.0
            0.0", rows=5, cols=1)
R = matrix("1.0 2.0 1.0 1.0", rows=1, cols=4)
IM = impurityMeasures(X = X, Y = Y, R = R, method = "entropy")

imputeByFD-Function

The imputeByFD-function imputes missing values from observed values (if exist) using robust functional dependencies.

Usage

imputeByFD(X, sourceAttribute, targetAttribute, threshold)

Arguments

Name Type Default Description
X Matrix[Double] Matrix of feature vectors (recoded matrix for non-numeric values)
source Integer Source attribute to use for imputation and error correction
target Integer Attribute to be fixed
threshold Double threshold value in interval [0, 1] for robust FDs

Returns

Type Description
Matrix[Double] Matrix with possible imputations

Example

X = matrix("1 1 1 2 4 5 5 3 3 NaN 4 5 4 1", rows=7, cols=2)
imputeByFD(X = X, source = 1, target = 2, threshold = 0.6, verbose = FALSE)

imputeEMA-Function

The imputeEMA-function imputes values with exponential moving average (single, double or triple).

Usage

ema(X, search_iterations, mode, freq, alpha, beta, gamma)

Arguments

Name Type Default Description
X Frame[Double] Frame that contains timeseries data that needs to be imputed
search_iterations Integer Budget iterations for parameter optimisation, used if parameters weren’t set
mode String Type of EMA method. Either “single”, “double” or “triple”
freq Double Seasonality when using triple EMA.
alpha Double alpha- value for EMA
beta Double beta- value for EMA
gamma Double gamma- value for EMA

Returns

Type Description
Frame[Double] Frame with EMA results

Example

X = read("fileA", data_type="frame")
ema(X = X, search_iterations = 1, mode = "triple", freq = 4, alpha = 0.1, beta = 0.1, gamma = 0.1,)

KMeans-Function

The kmeans() implements the KMeans Clustering algorithm.

Usage

kmeans(X = X, k = 20, runs = 10, max_iter = 5000, eps = 0.000001, is_verbose = FALSE, avg_sample_size_per_centroid = 50, seed = -1)

Arguments

Name Type Default Description
x Matrix[Double] required The input Matrix to do KMeans on.
k Int 10 Number of centroids
runs Int 10 Number of runs (with different initial centroids)
max_iter Int 100 Max no. of iterations allowed
eps Double 0.000001 Tolerance (epsilon) for WCSS change ratio
is_verbose Boolean FALSE do not print per-iteration stats
avg_sample_size_per_centroid int 50 Number of samples to make in the initialization
seed int -1 The seed used for initial sampling. If set to -1 random seeds are selected.

Returns

Type Description
String The mapping of records to centroids
String The output matrix with the centroids

Example

X = rand (rows = 3972, cols = 972)
kmeans(X = X, k = 20, runs = 10, max_iter = 5000, eps = 0.000001, is_verbose = FALSE, avg_sample_size_per_centroid = 50, seed = -1)

KNN-Function

The knn() implements the KNN (K Nearest Neighbor) algorithm.

Usage

[NNR, PR, FI] = knn(Train, Test, CL, k_value)

Arguments

Name Type Default Description
Train Matrix required The input matrix as features
Test Matrix required Number of centroids
CL Matrix Optional The input matrix as target
CL_T Integer 0 The target type of matrix CL whether columns in CL are continuous ( =1 ) or categorical ( =2 ) or not specified ( =0 )
trans_continuous Boolean FALSE Whether to transform continuous features to [-1,1]
k_value int 5 k value for KNN, ignore if select_k enable
select_k Boolean FALSE Use k selection algorithm to estimate k ( TRUE means yes )
k_min int 1 Min k value( available if select_k = 1 )
k_max int 100 Max k value( available if select_k = 1 )
select_feature Boolean FALSE Use feature selection algorithm to select feature ( TRUE means yes )
feature_max int 10 Max feature selection
interval int 1000 Interval value for K selecting ( available if select_k = 1 )
feature_importance Boolean FALSE Use feature importance algorithm to estimate each feature ( TRUE means yes )
predict_con_tg int 0 Continuous target predict function: mean(=0) or median(=1)
START_SELECTED Matrix Optional feature selection initial value

Returns

Type Description
Matrix NNR
Matrix PR
Matrix Feature importance value

Example

X = rand(rows = 100, cols = 20)
T = rand(rows= 3, cols = 20) # query rows, and columns
CL = matrix(seq(1,100), 100, 1)
k = 3
[NNR, PR, FI] = knn(Train=X, Test=T, CL=CL, k_value=k, predict_con_tg=1)

lenetTrain-Function

The lenetTrain-function trains LeNet CNN. The architecture of the networks is: conv1 -> relu1 -> pool1 -> conv2 -> relu2 -> pool2 -> affine3 -> relu3 -> affine4 -> softmax

Usage

model = lenetTrain(images, labels, images_val, labels_val, C, Hin, Win, 128, 3, 0.007, 0.9, 0.95, 5e-04, TRUE, -1)

Arguments

Name Type Default Description
X Matrix[Double] required Input data matrix, of shape (N, C*Hin*Win)
Y Matrix[Double] required Target matrix, of shape (N, K.)
X_val Matrix[Double] required Validation data matrix, of shape (N, C*Hin*Win)
Y_val Matrix[Double] required Validation target matrix, of shape (N, K)
C Integer required Number of input channels (dimensionality of input depth)
Win Integer required Input width
Hin Integer required Input height
batch_size Integer 64 Batch size
epochs Integer 20 Number of epochs
lr Double 0.01 Learning rate
mu Double 0.9 Momentum value
decay Double 0.95 Learning rate decay
lambda Double 5e-04 Regularization strength
seed Integer -1 Seed for model initialization. Seed value of -1 will generate random seed.
verbose Boolean FALSE Flag indicates if function should print to stdout

Returns

Type Description
list[unknown] Trained model which can be used in lenetPredict.

Example

data = read(path, format="csv")
images = images = train_data[,2:ncol(data)]
labels_int = data[,1]
C = 1
Hin = 28
Win = 28

# Scale images to [-1,1], and one-hot encode the labels
n = nrow(train_data)
images = (images / 255.0) * 2 - 1
labels = table(seq(1, n), labels_int+1, n, 10)

model = lenetTrain(images, labels, images_val, labels_val, C, Hin, Win, 128, 3, 0.007, 0.9, 0.95, 5e-04, TRUE, -1)

lenetPredict-Function

The lenetPredict-function makes prediction given the data and trained LeNet model.

Usage

probs = lenetPredict(model=model, X=images, C=C, Hin=Hin, Win=Win)

Arguments

Name Type Default Description
model list[unknown] required Trained LeNet model
X Matrix[Double] required Input data matrix, of shape (N, C*Hin*Win)
C Integer required Number of input channels (dimensionality of input depth)
Win Integer required Input width
Hin Integer required Input height
batch_size Integer 64 Batch size

Returns

Type Description
Matrix[Double] Predicted values

Example

model = lenetTrain(images, labels, images_val, labels_val, C, Hin, Win, 128, 3, 0.007, 0.9, 0.95, 5e-04, TRUE, -1)
probs = lenetPredict(model=model, X=images_test, C=C, Hin=Hin, Win=Win)

lm-Function

The lm-function solves linear regression using either the direct solve method or the conjugate gradient algorithm depending on the input size of the matrices (See lmDS-function and lmCG-function respectively).

Usage

lm(X, y, icpt = 0, reg = 1e-7, tol = 1e-7, maxi = 0, verbose = TRUE)

Arguments

Name Type Default Description
X Matrix[Double] required Matrix of feature vectors.
y Matrix[Double] required 1-column matrix of response values.
icpt Integer 0 Intercept presence, shifting and rescaling the columns of X (Details)
reg Double 1e-7 Regularization constant (lambda) for L2-regularization. set to nonzero for highly dependant/sparse/numerous features
tol Double 1e-7 Tolerance (epsilon); conjugate gradient procedure terminates early if L2 norm of the beta-residual is less than tolerance * its initial norm
maxi Integer 0 Maximum number of conjugate gradient iterations. 0 = no maximum
verbose Boolean TRUE If TRUE print messages are activated

Note that if number of features is small enough (rows of X/y < 2000), the lmDS-Function’ is called internally and parameters tol and maxi are ignored.

Returns

Type Description
Matrix[Double] 1-column matrix of weights.
icpt-Argument

The icpt-argument can be set to 3 modes:

Example

X = rand (rows = 50, cols = 10)
y = X %*% rand(rows = ncol(X), cols = 1)
lm(X = X, y = y)

intersect-Function

The intersect-function implements set intersection for numeric data.

Usage

intersect(X, Y)

Arguments

Name Type Default Description
X Double matrix X, set A
Y Double matrix Y, set B

Returns

Type Description
Double intersection matrix, set of intersecting items

lmCG-Function

The lmCG-function solves linear regression using the conjugate gradient algorithm.

Usage

lmCG(X, y, icpt = 0, reg = 1e-7, tol = 1e-7, maxi = 0, verbose = TRUE)

Arguments

Name Type Default Description
X Matrix[Double] required Matrix of feature vectors.
y Matrix[Double] required 1-column matrix of response values.
icpt Integer 0 Intercept presence, shifting and rescaling the columns of X (Details)
reg Double 1e-7 Regularization constant (lambda) for L2-regularization. set to nonzero for highly dependant/sparse/numerous features
tol Double 1e-7 Tolerance (epsilon); conjugate gradient procedure terminates early if L2 norm of the beta-residual is less than tolerance * its initial norm
maxi Integer 0 Maximum number of conjugate gradient iterations. 0 = no maximum
verbose Boolean TRUE If TRUE print messages are activated

Returns

Type Description
Matrix[Double] 1-column matrix of weights.

Example

X = rand (rows = 50, cols = 10)
y = X %*% rand(rows = ncol(X), cols = 1)
lmCG(X = X, y = y, maxi = 10)

lmDS-Function

The lmDS-function solves linear regression by directly solving the linear system.

Usage

lmDS(X, y, icpt = 0, reg = 1e-7, verbose = TRUE)

Arguments

Name Type Default Description
X Matrix[Double] required Matrix of feature vectors.
y Matrix[Double] required 1-column matrix of response values.
icpt Integer 0 Intercept presence, shifting and rescaling the columns of X (Details)
reg Double 1e-7 Regularization constant (lambda) for L2-regularization. set to nonzero for highly dependant/sparse/numerous features
verbose Boolean TRUE If TRUE print messages are activated

Returns

Type Description
Matrix[Double] 1-column matrix of weights.

Example

X = rand (rows = 50, cols = 10)
y = X %*% rand(rows = ncol(X), cols = 1)
lmDS(X = X, y = y)

lmPredict-Function

The lmPredict-function predicts the class of a feature vector.

Usage

lmPredict(X=X, B=w, ytest= Y)

Arguments

Name Type Default Description
X Matrix[Double] required Matrix of feature vector(s).
B Matrix[Double] required 1-column matrix of weights.
ytest Matrix[Double] required test labels, used only for verbose output. can be set to matrix(0,1,1) if verbose output is not wanted
icpt Integer 0 Intercept presence, shifting and rescaling of X (Details)
verbose Boolean FALSE Print various statistics for evaluating accuracy.

Returns

Type Description
Matrix[Double] 1-column matrix of classes.

Example

X = rand (rows = 50, cols = 10)
y = X %*% rand(rows = ncol(X), cols = 1)
w = lm(X = X, y = y)
yp = lmPredict(X = X, B = w, ytest=matrix(0,1,1))

matrixProfile-Function

The matrixProfile-function implements the SCRIMP algorithm for efficient time-series analysis.

Usage

matrixProfile(ts, window_size, sample_percent, is_verbose)

Arguments

| Name | Type | Default | Description | | :—— | :————- | ——– | :———- | | ts | Matrix | — | Input Frame X | | window_size | Integer | 4 | Sliding window size | | sample_percent| Double | 1.0 | Degree of approximation between zero and one (1 computes the exact solution) | | verbose | Boolean | False | Print debug information |

Returns

Type Default Description
Matrix[Double] The computed matrix profile distances
Matrix[Integer] Indices of least distances

mdedup-Function

The mdedup-function implements builtin for deduplication using matching dependencies (e.g. Street 0.95, City 0.90 -> ZIP 1.0) by Jaccard distance.

Usage

mdedup(X, Y, intercept, epsilon, lamda, maxIterations, verbose)

Arguments

Name Type Default Description
X Frame Input Frame X
LHSfeatures Matrix[Integer] A matrix 1xd with numbers of columns for MDs
LHSthreshold Matrix[Double] A matrix 1xd with threshold values in interval [0, 1] for MDs
RHSfeatures Matrix[Integer] A matrix 1xd with numbers of columns for MDs
RHSthreshold Matrix[Double] A matrix 1xd with threshold values in interval [0, 1] for MDs
verbose Boolean False Set to true to print duplicates.

Returns

Type Default Description
Matrix[Integer] Matrix of duplicates (rows).

Example

X = as.frame(rand(rows = 50, cols = 10))
LHSfeatures = matrix("1 3 19", 1, 2)
LHSthreshold = matrix("0.85 0.85", 1, 2)
RHSfeatures = matrix("30", 1, 1)
RHSthreshold = matrix("1.0", 1, 1)
duplicates = mdedup(X, LHSfeatures, LHSthreshold, RHSfeatures, RHSthreshold, verbose = FALSE)

mice-Function

The mice-function implements Multiple Imputation using Chained Equations (MICE) for nominal data.

Usage

mice(F, cMask, iter, complete, verbose)

Arguments

Name Type Default Description
X Matrix[Double] required Data Matrix (Recoded Matrix for categorical features), ncol(X) > 1
cMask Matrix[Double] required 0/1 row vector for identifying numeric (0) and categorical features (1) with one-dimensional row matrix with column = ncol(F).
iter Integer 3 Number of iteration for multiple imputations.
verbose Boolean FALSE Boolean value.

Returns

Type Description
Matrix[Double] imputed dataset.

Example

F = matrix("4 3 NaN 8 7 8 5 NaN 6", rows=3, cols=3)
cMask = round(rand(rows=1,cols=ncol(F),min=0,max=1))
dataset = mice(F, cMask, iter = 3, verbose = FALSE)

msvm-Function

The msvm-function implements builtin multiclass SVM with squared slack variables It learns one-against-the-rest binary-class classifiers by making a function call to l2SVM

Usage

msvm(X, Y, intercept, epsilon, lamda, maxIterations, verbose)

Arguments

Name Type Default Description
X Double Matrix X of feature vectors.
Y Double Matrix Y of class labels.
intercept Boolean False No Intercept ( If set to TRUE then a constant bias column is added to X)
num_classes Integer 10 Number of classes.
epsilon Double 0.001 Procedure terminates early if the reduction in objective function value is less than epsilon (tolerance) times the initial objective function value.
lamda Double 1.0 Regularization parameter (lambda) for L2 regularization
maxIterations Integer 100 Maximum number of conjugate gradient iterations
verbose Boolean False Set to true to print while training.

Returns

Name Type Default Description
model Double Model matrix.

Example

X = rand(rows = 50, cols = 10)
y = round(X %*% rand(rows=ncol(X), cols=1))
model = msvm(X = X, Y = y, intercept = FALSE, epsilon = 0.005, lambda = 1.0, maxIterations = 100, verbose = FALSE)

multiLogReg-Function

The multiLogReg-function solves Multinomial Logistic Regression using Trust Region method. (See: Trust Region Newton Method for Logistic Regression, Lin, Weng and Keerthi, JMLR 9 (2008) 627-650)

Usage

multiLogReg(X, Y, icpt, reg, tol, maxi, maxii, verbose)

Arguments

Name Type Default Description
X Double The matrix of feature vectors
Y Double The matrix with category labels
icpt Int 0 Intercept presence, shifting and rescaling X columns: 0 = no intercept, no shifting, no rescaling; 1 = add intercept, but neither shift nor rescale X; 2 = add intercept, shift & rescale X columns to mean = 0, variance = 1
reg Double 0 regularization parameter (lambda = 1/C); intercept is not regularized
tol Double 1e-6 tolerance (“epsilon”)
maxi Int 100 max. number of outer newton interations
maxii Int 0 max. number of inner (conjugate gradient) iterations

Returns

Type Description
Double Regression betas as output for prediction

Example

X = rand(rows = 50, cols = 30)
Y = X %*% rand(rows = ncol(X), cols = 1)
betas = multiLogReg(X = X, Y = Y, icpt = 2,  tol = 0.000001, reg = 1.0, maxi = 100, maxii = 20, verbose = TRUE)

naiveBayes-Function

The naiveBayes-function computes the class conditional probabilities and class priors.

Usage

naiveBayes(D, C, laplace, verbose)

Arguments

Name Type Default Description
D Matrix[Double] required One dimensional column matrix with N rows.
C Matrix[Double] required One dimensional column matrix with N rows.
Laplace Double 1 Any Double value.
Verbose Boolean TRUE Boolean value.

Returns

Type Description
Matrix[Double] Class priors, One dimensional column matrix with N rows.
Matrix[Double] Class conditional probabilites, One dimensional column matrix with N rows.

Example

D=rand(rows=10,cols=1,min=10)
C=rand(rows=10,cols=1,min=10)
[prior, classConditionals] = naiveBayes(D, C, laplace = 1, verbose = TRUE)

naiveBaysePredict-Function

The naiveBaysePredict-function predicts the scoring with a naive Bayes model.

Usage

naiveBaysePredict(X=X, P=P, C=C)

Arguments

Name Type Default Description
X Matrix[Double] required Matrix of test data with N rows.
P Matrix[Double] required Class priors, One dimensional column matrix with N rows.
C Matrix[Double] required Class conditional probabilities, matrix with N rows.

Returns

Type Description
Matrix[Double] A matrix containing the top-K item-ids with highest predicted ratings.
Matrix[Double] A matrix containing predicted ratings.

Example

[YRaw, Y] = naiveBaysePredict(X=data, P=model_prior, C=model_conditionals)

normalize-Function

The normalize-function normalises the values of a matrix by changing the dataset to use a common scale. This is done while preserving differences in the ranges of values. The output is a matrix of values in range [0,1].

Usage

normalize(X); 

Arguments

Name Type Default Description
X Matrix[Double] required Matrix of feature vectors.

Returns

Type Description
Matrix[Double] 1-column matrix of normalized values.

Example

X = rand(rows = 50, cols = 10)
y = X %*% rand(rows = ncol(X), cols = 1)
y = normalize(X = X)

outlier-Function

This outlier-function takes a matrix data set as input from where it determines which point(s) have the largest difference from mean.

Usage

outlier(X, opposite)

Arguments

Name Type Default Description
X Matrix[Double] required Matrix of Recoded dataset for outlier evaluation
opposite Boolean required (1)TRUE for evaluating outlier from upper quartile range, (0)FALSE for evaluating outlier from lower quartile range

Returns

Type Description
Matrix[Double] matrix indicating outlier values

Example

X = rand (rows = 50, cols = 10)
outlier(X=X, opposite=1)

outlierByDB-Function

The outlierByDB-function implements an outlier prediction for a trained dbscan model. The points in the Xtest matrix are checked against the model and are considered part of the cluster if at least one member is within eps distance.

Usage

outlierByDB(X, model, eps)

Arguments

Name Type Default Description
Xtest Matrix[Double] required Matrix of points for outlier testing
model Matrix[Double] required Matrix model of the clusters, containing all points that are considered members, returned by the dbscan builtin
eps Double 0.5 Epsilon distance between points to be considered in their neighborhood

Returns

Type Description
Matrix[Double] Matrix indicating outlier values of the points in Xtest, 0 suggests it being an outlier

Example

eps = 1
minPts = 5
X = rand(rows=1780, cols=180, min=1, max=20)
[indices, model] = dbscan(X=X, eps=eps, minPts=minPts)
Y = rand(rows=500, cols=180, min=1, max=20)
Z = outlierByDB(Xtest=Y, clusterModel = model, eps = eps)

pnmf-Function

The pnmf-function implements Poisson Non-negative Matrix Factorization (PNMF). Matrix X is factorized into two non-negative matrices, W and H based on Poisson probabilistic assumption. This non-negativity makes the resulting matrices easier to inspect.

Usage

pnmf(X, rnk, eps = 10^-8, maxi = 10, verbose = TRUE)

Arguments

Name Type Default Description
X Matrix[Double] required Matrix of feature vectors.
rnk Integer required Number of components into which matrix X is to be factored.
eps Double 10^-8 Tolerance
maxi Integer 10 Maximum number of conjugate gradient iterations.
verbose Boolean TRUE If TRUE, ‘iter’ and ‘obj’ are printed.

Returns

Type Description
Matrix[Double] List of pattern matrices, one for each repetition.
Matrix[Double] List of amplitude matrices, one for each repetition.

Example

X = rand(rows = 50, cols = 10)
[W, H] = pnmf(X = X, rnk = 2, eps = 10^-8, maxi = 10, verbose = TRUE)

scale-Function

The scale function is a generic function whose default method centers or scales the column of a numeric matrix.

Usage

scale(X, center=TRUE, scale=TRUE)

Arguments

Name Type Default Description
X Matrix[Double] required Matrix of feature vectors.
center Boolean required either a logical value or numerical value.
scale Boolean required either a logical value or numerical value.

Returns

Type Description
Matrix[Double] 1-column matrix of weights.

Example

X = rand(rows = 20, cols = 10)
center=TRUE;
scale=TRUE;
Y= scale(X,center,scale)

setdiff-Function

The setdiff-function returns the values of X that are not in Y.

Usage

setdiff(X, Y)

Arguments

Name Type Default Description
X Matrix[Double] required input vector
Y Matrix[Double] required input vector

Returns

Type Description
Matrix[Double] values of X that are not in Y.

Example

X = matrix("1 2 3 4", rows = 4, cols = 1)
Y = matrix("2 3", rows = 2, cols = 1)
R = setdiff(X = X, Y = Y)

sherlock-Function

Implements training phase of Sherlock: A Deep Learning Approach to Semantic Data Type Detection

[Hulsebos, Madelon, et al. “Sherlock: A deep learning approach to semantic data type detection.” Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining., 2019]

Usage

sherlock(X_train, y_train)

Arguments

Name Type Default Description
X_train Matrix[Double] required Matrix of feature vectors.
y_train Matrix[Double] required Matrix Y of class labels of semantic data type.

Returns

Type Description
Matrix[Double] weights (parameters) matrices for character distribtions
Matrix[Double] weights (parameters) matrices for word embeddings
Matrix[Double] weights (parameters) matrices for paragraph vectors
Matrix[Double] weights (parameters) matrices for global statistics
Matrix[Double] weights (parameters) matrices for combining all featurs (final)

Example

# preprocessed training data taken from sherlock corpus
processed_train_values = read("processed/X_train.csv")
processed_train_labels = read("processed/y_train.csv")
transform_spec = read("processed/transform_spec.json")
[processed_train_labels, label_encoding] = sherlock::transform_encode_labels(processed_train_labels, transform_spec)
write(label_encoding, "weights/label_encoding")

[cW1,  cb1,
 cW2,  cb2,
 cW3,  cb3,
 wW1,  wb1,
 wW2,  wb2,
 wW3,  wb3,
 pW1,  pb1,
 pW2,  pb2,
 pW3,  pb3,
 sW1,  sb1,
 sW2,  sb2,
 sW3,  sb3,
 fW1,  fb1,
 fW2,  fb2,
 fW3,  fb3] = sherlock(processed_train_values, processed_train_labels)

sherlockPredict-Function

Implements prediction and evaluation phase of Sherlock: A Deep Learning Approach to Semantic Data Type Detection

[Hulsebos, Madelon, et al. “Sherlock: A deep learning approach to semantic data type detection.” Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining., 2019]

Usage

sherlockPredict(X, cW1, cb1, cW2, cb2, cW3, cb3, wW1, wb1, wW2, wb2, wW3, wb3,
                   pW1, pb1, pW2, pb2, pW3, pb3, sW1, sb1, sW2, sb2, sW3, sb3,
                   fW1, fb1, fW2, fb2, fW3, fb3)

Arguments

Name Type Default Description
X Matrix[Double] required Matrix of values which are to be classified.
cW Matrix[Double] required Weights (parameters) matrices for character distribtions.
cb Matrix[Double] required Biases vectors for character distribtions.
wW Matrix[Double] required Weights (parameters) matrices for word embeddings.
wb Matrix[Double] required Biases vectors for word embeddings.
pW Matrix[Double] required Weights (parameters) matrices for paragraph vectors.
pb Matrix[Double] required Biases vectors for paragraph vectors.
sW Matrix[Double] required Weights (parameters) matrices for global statistics.
sb Matrix[Double] required Biases vectors for global statistics.
fW Matrix[Double] required Weights (parameters) matrices combining all features (final).
fb Matrix[Double] required Biases vectors combining all features (final).

Returns

Type Description
Matrix[Double] Class probabilities of shape (N, K).

Example

# preprocessed validation data taken from sherlock corpus
processed_val_values = read("processed/X_val.csv")
processed_val_labels = read("processed/y_val.csv")
transform_spec = read("processed/transform_spec.json")
label_encoding = read("weights/label_encoding")
processed_val_labels = sherlock::transform_apply_labels(processed_val_labels, label_encoding, transform_spec)

probs = sherlockPredict(processed_val_values, cW1,  cb1,
cW2,  cb2,
cW3,  cb3,
wW1,  wb1,
wW2,  wb2,
wW3,  wb3,
pW1,  pb1,
pW2,  pb2,
pW3,  pb3, 
sW1,  sb1,
sW2,  sb2,
sW3,  sb3,
fW1,  fb1,
fW2,  fb2,
fW3,  fb3)

[loss, accuracy] = sherlockPredict::eval(probs, processed_val_labels)

sigmoid-Function

The Sigmoid function is a type of activation function, and also defined as a squashing function which limit the output to a range between 0 and 1, which will make these functions useful in the prediction of probabilities.

Usage

sigmoid(X)

Arguments

Name Type Default Description
X Matrix[Double] required Matrix of feature vectors.

Returns

Type Description
Matrix[Double] 1-column matrix of weights.

Example

X = rand (rows = 20, cols = 10)
Y = sigmoid(X)

slicefinder-Function

The slicefinder-function returns top-k worst performing subsets according to a model calculation.

Usage

slicefinder(X,W, y, k, paq, S);

Arguments

Name Type Default Description
X Matrix[Double] required Recoded dataset into Matrix
W Matrix[Double] required Trained model
y Matrix[Double] required 1-column matrix of response values.
k Integer 1 Number of subsets required
paq Integer 1 amount of values wanted for each col, if paq = 1 then its off
S Integer 2 amount of subsets to combine (for now supported only 1 and 2)

Returns

Type Description
Matrix[Double] Matrix containing the information of top_K slices (relative error, standart error, value0, value1, col_number(sort), rows, cols,range_row,range_cols, value00, value01,col_number2(sort), rows2, cols2,range_row2,range_cols2)

Usage

X = rand (rows = 50, cols = 10)
y = X %*% rand(rows = ncol(X), cols = 1)
w = lm(X = X, y = y)
ress = slicefinder(X = X,W = w, Y = y,  k = 5, paq = 1, S = 2);

smote-Function

The smote-function (Synthetic Minority Oversampling Technique) implements a classical techniques for handling class imbalance. The built-in takes the samples from minority class and over-sample them by generating the synthesized samples. The built-in accepts two parameters s and k. The parameter s define the number of synthesized samples to be generated i.e., over-sample the minority class by s time, where s is the multiple of 100. For given 40 samples of minority class and s = 200 the smote will generate the 80 synthesized samples to over-sample the class by 200 percent. The parameter k is used to generate the k nearest neighbours for each minority class sample and then the neighbours are chosen randomly in synthesis process.

Usage

smote(X, s, k, verbose);

Arguments

Name Type Default Description
X Matrix[Double] required Matrix of feature vectors of minority class samples
s Integer 200 Amount of SMOTE (percentage of oversampling), integral multiple of 100
k Integer 1 Number of nearest neighbour
verbose Boolean TRUE If TRUE print messages are activated

Returns

Type Description
Matrix[Double] Matrix of (N/100) * X synthetic minority class samples

Example

X = rand (rows = 50, cols = 10)
B = smote(X = X, s=200, k=3, verbose=TRUE);

steplm-Function

The steplm-function (stepwise linear regression) implements a classical forward feature selection method. This method iteratively runs what-if scenarios and greedily selects the next best feature until the Akaike information criterion (AIC) does not improve anymore. Each configuration trains a regression model via lm, which in turn calls either the closed form lmDS or iterative lmGC.

Usage

steplm(X, y, icpt);

Arguments

Name Type Default Description
X Matrix[Double] required Matrix of feature vectors.
y Matrix[Double] required 1-column matrix of response values.
icpt Integer 0 Intercept presence, shifting and rescaling the columns of X (Details)
reg Double 1e-7 Regularization constant (lambda) for L2-regularization. set to nonzero for highly dependent/sparse/numerous features
tol Double 1e-7 Tolerance (epsilon); conjugate gradient procedure terminates early if L2 norm of the beta-residual is less than tolerance * its initial norm
maxi Integer 0 Maximum number of conjugate gradient iterations. 0 = no maximum
verbose Boolean TRUE If TRUE print messages are activated

Returns

Type Description
Matrix[Double] Matrix of regression parameters (the betas) and its size depend on icpt input value. (C in the example)
Matrix[Double] Matrix of selected features ordered as computed by the algorithm. (S in the example)
icpt-Argument

The icpt-arg can be set to 2 modes:

selected-Output

If the best AIC is achieved without any features the matrix of selected features contains 0. Moreover, in this case no further statistics will be produced

Example

X = rand (rows = 50, cols = 10)
y = X %*% rand(rows = ncol(X), cols = 1)
[C, S] = steplm(X = X, y = y, icpt = 1);

symmetricDifference-Function

The symmetricDifference-function returns the symmetric difference of the two input vectors. This is done by calculating the setdiff (nonsymmetric) between union and intersect of the two input vectors.

Usage

symmetricDifference(X, Y)

Arguments

Name Type Default Description
X Matrix[Double] required input vector
Y Matrix[Double] required input vector

Returns

Type Description
Matrix[Double] symmetric difference of the input vectors

Example

X = matrix("1 2 3.1", rows = 3, cols = 1)
Y = matrix("3.1 4", rows = 2, cols = 1)
R = symmetricDifference(X = X, Y = Y)

The tomekLink-function performs undersampling by removing Tomek’s links for imbalanced multiclass problems

Reference: “Two Modifications of CNN,” in IEEE Transactions on Systems, Man, and Cybernetics, vol. SMC-6, no. 11, pp. 769-772, Nov. 1976, doi: 10.1109/TSMC.1976.4309452.

Usage

[X_under, y_under, drop_idx] = tomeklink(X, y)

Arguments

Name Type Default Description
X Matrix[Double] required Data Matrix (n,m)
y Matrix[Double] required Label Matrix (n,1)

Returns

Name Type Description
X_under Matrix[Double] Data Matrix without Tomek links
y_under Matrix[Double] Labels corresponding to undersampled data
drop_idx Matrix[Double] Indices of dropped rows/labels wrt. input

Example

X = round(rand(rows = 53, cols = 6, min = -1, max = 1))
y = round(rand(rows = nrow(X), cols = 1, min = 0, max = 1))
[X_under, y_under, drop_idx] = tomeklink(X, y)

toOneHot-Function

The toOneHot-function encodes unordered categorical vector to multiple binarized vectors.

Usage

toOneHot(X, numClasses)

Arguments

Name Type Default Description
X Matrix[Double] required vector with N integer entries between 1 and numClasses.
numClasses int required number of columns, must be greater than or equal to largest value in X.

Returns

Type Description
Matrix[Double] one-hot-encoded matrix with shape (N, numClasses).

Example

numClasses = 5
X = round(rand(rows = 10, cols = 10, min = 1, max = numClasses))
y = toOneHot(X,numClasses)

tSNE-Function

The tSNE-function performs dimensionality reduction using tSNE algorithm based on the paper: Visualizing Data using t-SNE, Maaten et. al.

Usage

tSNE(X, reduced_dims, perplexity, lr, momentum, max_iter, seed, is_verbose)

Arguments

| Name | Type | Default | Description | | :———– | :————- | ——– | :———- | | X | Matrix[Double] | required | Data Matrix of shape (number of data points, input dimensionality) | | reduced_dims | Integer | 2 | Output dimensionality | | perplexity | Integer | 30 | Perplexity Parameter | | lr | Double | 300. | Learning rate | | momentum | Double | 0.9 | Momentum Parameter | | max_iter | Integer | 1000 | Number of iterations | | seed | Integer | -1 | The seed used for initial values. If set to -1 random seeds are selected. | | is_verbose | Boolean | FALSE | Print debug information |

Returns

Type Description
Matrix[Double] Data Matrix of shape (number of data points, reduced_dims)

Example

X = rand(rows = 100, cols = 10, min = -10, max = 10))
Y = tSNE(X)

union-Function

The union-function combines all rows from both input vectors and removes all duplicate rows by calling unique on the resulting vector.

Usage

union(X, Y)

Arguments

Name Type Default Description
X Matrix[Double] required input vector
Y Matrix[Double] required input vector

Returns

Type Description
Matrix[Double] the union of both input vectors.

Example

X = matrix("1 2 3 4", rows = 4, cols = 1)
Y = matrix("3 4 5 6", rows = 4, cols = 1)
R = union(X = X, Y = Y)

unique-Function

The unique-function returns a set of unique rows from a given input vector.

Usage

unique(X)

Arguments

Name Type Default Description
X Matrix[Double] required input vector

Returns

Type Description
Matrix[Double] a set of unique values from the input vector

Example

X = matrix("1 3.4 7 3.4 -0.9 8 1", rows = 7, cols = 1)
R = unique(X = X)

winsorize-Function

The winsorize-function removes outliers from the data. It does so by computing upper and lower quartile range of the given data then it replaces any value that falls outside this range (less than lower quartile range or more than upper quartile range).

Usage

winsorize(X)

Arguments

Name Type Default Description
X Matrix[Double] required recorded data set with possible outlier values

Returns

Type Description
Matrix[Double] Matrix without outlier values

Example

X = rand(rows=10, cols=10,min = 1, max=9)
Y = winsorize(X=X)

xgboost-Function

XGBoost is a decision-tree-based ensemble Machine Learning algorithm that uses a gradient boosting. This xgboost implementation supports classification and regression and is capable of working with categorical and scalar features.

Usage

M = xgboost(X = X, y = y, R = R, sml_type = 1, num_trees = 3, learning_rate = 0.3, max_depth = 6, lambda = 0.0)

Arguments

NAME TYPE DEFAULT Description
X Matrix[Double] Feature matrix X; categorical features needs to be one-hot-encoded
y Matrix[Double] Label matrix y
R Matrix[Double] Matrix R; 1xn vector which for each feature in X contains the following information
      - R[,2]: 1 (scalar feature)
      - R[,1]: 2 (categorical feature)
sml_type Integer 1 Supervised machine learning type: 1 = Regression(default), 2 = Classification
num_trees Integer 10 Number of trees to be created in the xgboost model
learning_rate Double 0.3 alias: eta. After each boosting step the learning rate controls the weights of the new predictions
max_depth Integer 6 Maximum depth of a tree. Increasing this value will make the model more complex and more likely to overfit
lambda Double 0.0 L2 regularization term on weights. Increasing this value will make model more conservative and reduce amount of leaves of a tree

Returns

| Name | Type | Default | Description | | :— | :————- | ——- | :———————————————————– | | M | Matrix[Double] | — | Each column of the matrix corresponds to a node in the learned model
Detailed description can be found in xgboost.dml |

Example

X = matrix("4.5 3.0 3.0 2.8 3.5
            1.9 2.0 1.0 3.4 2.9
            2.0 1.0 1.0 4.9 3.4
            2.3 2.0 2.0 1.4 1.8
            2.1 1.0 3.0 1.0 1.9", rows=5, cols=5)
Y = matrix("1.0
            4.0
            4.0
            7.0
            8.0", rows=5, cols=1)
R = matrix("1.0 1.0 1.0 1.0 1.0", rows=1, cols=5)
M = xgboost(X = X, y = Y, R = R)

xgboostPredict-Function

In order to calculate a prediction, XGBoost sums predictions of all its trees. Each tree is not a great predictor on it’s own, but by summing across all trees, XGBoost is able to provide a robust prediction in many cases. Depending on our supervised machine learning type use xgboostPredictRegression() or xgboostPredictClassification() to predict the labels.

Usage

y_pred = xgboostPredictRegression(X = X, M = M)

or

y_pred = xgboostPredictClassification(X = X, M = M)

Arguments

NAME TYPE DEFAULT Description
X Matrix[Double] Feature matrix X; categorical features needs to be one-hot-encoded
M Matrix[Double] Trained model returned from xgboost. Each column of the matrix corresponds to a node in the learned model
Detailed description can be found in xgboost.dml
learning_rate Double 0.3 alias: eta. After each boosting step the learning rate controls the weights of the new predictions. Should be the same as at xgboost-function call

Returns

Name Type Default Description
P Matrix[Double] xgboostPredictRegression: The prediction of the samples using the xgboost model. (y_prediction)
xgboostPredictClassification: The probability of the samples being 1 (like XGBClassifier.predict_proba() in Python)

Example

X = matrix("4.5 3.0 3.0 2.8 3.5
            1.9 2.0 1.0 3.4 2.9
            2.0 1.0 1.0 4.9 3.4
            2.3 2.0 2.0 1.4 1.8
            2.1 1.0 3.0 1.0 1.9", rows=5, cols=5)
Y = matrix("1.0
            4.0
            4.0
            7.0
            8.0", rows=5, cols=1)
R = matrix("1.0 1.0 1.0 1.0 1.0", rows=1, cols=5)
M = xgboost(X = X, y = Y, R = R, num_trees = 10, learning_rate = 0.4)
P = xgboostPredictRegression(X = X, M = M, learning_rate = 0.4)