Feature selection for hard voting classifier and NN sparse weight initialization.
Generate different types of sparsity pattern for sparse matrices.
collection of utility functions for correlation analysis
Feature selection for Hard Voting classifier
Split strings into (character-based) k-shingles
Sentence embedding evaluation for German
Convert W2V embeddings of a sequence (2D) to one vector (1D)
Batch Correlation Regularizer for TF2/Keras
Add a regularization if the features/columns/neurons the hidden layer or output layer should be correlated. The vector with target correlation coefficient is computed before the optimization, and compared with correlation coefficients computed across the batch examples.
Input Embedding Training as Similarity Learning Problem (SimiVec)
Equilibrated Input Embedding Initialization (EIEI)
Solving quadratic optimization problems with Keras.
Training of multi-label embeddings for k-shingled input sequences for PyTorch.
Training of multi-label embeddings for k-shingled input sequences. for Tensorflow2/Keras
Applying declination and conjugation rules to lemmata.
Lagmatrix. Create array with time-lagged copies of the features
Compute similarity between netsted set based trees.
Compute distance between all nodes of a tree, and estimate an histogram that can be used as features for other models.
Properties of IPA symbols for data analysis.
Daycount methods to compute date differences in year units
Sampling algorithm for best-worst scaling sets.
PyTorch implementation of transformer block layers.
Wrapper and utility functions to apply scipy's SLSQP algorithm to quadratic optimization problems with resource constraints and upper boundaries.
Boilerplate code to wrap different libs for NLP tasks.
Fractional Difference for Time Series
Three-way data split into training set, validation set, and test set.
Utility functions for Keras/Tensorflow2.
Utility functions for PyTorch.
sklearn wrapper for numpy-fracdiff
Utility functions for scipy.
Generate a list of random dates or resp. datetime objects
pad variable length sequences with multiples features
Linear Regression with numpy only.
Determine fractal order by the ADF test
Feature engineering sklearn transformer for dates
save and load API keys from a file
transform an ill-conditioned quadratic matrix to a positive semidefinite matrix
deprecated. Please use "keras_tweaks.dense_sparse_matmul" instead.
deprecated. Please use "keras_tweaks.get_sparsity_pattern" instead.
additional wrapper classes for the sklearn API
pandas utility functions
RNNs with layer normalization
Some utility function for the numerai competition
utility functions for numpy
Encode grouped labels
Tweaks for jupyter notebooks
One-Hot encoder with sklearn-ish API interface that process mixed string and numeric labels directly.
Thermometer Encoding for ordinal variables
Continous Time Markov Chain
merge probability predictions for majority votes
Scale a variable into an open interval (0,1) whereas values within a given lower and upper percentile maintain a linear relation, and outlier saturate towards the interval limits.
model zoo of different preconfigured algorithms
Continuous Time Markov Chain for daily panel data and annual transition probabilities
classes and functions to download or scrape data
Tweaks for Google Colab
Jackknife resampling, parameter estimation and stability test
Continous Time Markov Chain (with automatic error correction)
(deprecated package) Markov Modeling