List of Availble Algorithms

Notation

To avoid confusion of terms and notations, we make the following basic definitions:

\(X\): feature matrix

\[\begin{split}X = \begin{pmatrix} x_{11} & x_{12} & \dots & x_{1m} \\ x_{21} & x_{22} & \dots & x_{2m} \\ \vdots & \vdots & \dots & \vdots \\ v_{n1} & x_{n2} & \dots & x_{nm} \\ \end{pmatrix}\end{split}\]

Here, each row denotes a sample (or an observation) \(x_i (i=1, \dots, n)\) or \(X_{i.}\) . Each column denotes a feature, \(X_{.j} (j = 1, \dots , p)\)

\(Y\): the label in supervised learning

\[\begin{split}Y = \begin{pmatrix} y_1\\ y_2\\ \vdots \\ y_n \\ \end{pmatrix}\end{split}\]

We have \(y_i \in \mathcal{R}\) for regression problem, and \(y_i \in \mathcal{Z}\) for classification

Training/Validation dataset: in XFL, we mainly use two types of dataset, one for training and the other for validation. We use the superscript “train”, “val” to identify them. For example, \(X^{train}\), \(X^{val}\) denote training dataset, validation dataset respectively.

List of Algorithms

Algorithm	Module	Description
Local Normalization	local/normalization	normalize data
Local Standard Scaler	local/standard_scaler	standardize data
Local Data Split	local/data_split	split data into train and validation set
Local Feature Preprocess	local/feature_preprocess	feature preprocess
Local Data Statistic	local/data_statistic	data statistic
Horizontal Linear Regression	horizontal/linear_regression	two-party or multi-party horizontal linear regression
Horizontal Logistic Regression	horizontal/logistic_regression	two-party or multi-party horizontal logistic regression
Horizontal Poisson Regression	horizontal/poisson_regression	two-party or multi-party horizontal poisson regression
Horizontal ResNet	horizontal/Resnet	two-party or multi-party horizontal ResNet
Horizontal DenseNet	horizontal/Densenet	two-party or multi-party horizontal DenseNet
Horizontal VGG	horizontal/Vgg	two-party or multi-party horizontal VGG
Horizontal Bert	horizontal/Bert	two-party or multi-party horizontal Bert
Vertical Binning Woe IV	vertical/binning_woe_iv	calulate WoE and IV using equal-frequency binning or equal-width binning
Vertical Pearson	vertical/pearson	two-party or multi-party vertical Pearson correlation coefficient
Vertical Feature Selection	vertical/feature_selection	two-party or multi-party vertical feature selection
Vertical Logistic Regression	vertical/logistic_regression	two-party or multi-party vertical logistic regression
Vertical Linear Regression	vertical/linear_regression	two-party or multi-party vertical linear regression
Vertical Poisson Regression	vertical/poisson_regression	two-party or multi-party vertical poisson regression
Vertical XGBoost	vertical/xgboost	two-party or multi-party vertical xgboost
Vertical XGBoostDistributed	vertical/xgboost_distributed	distributed two-party or multi-party vertical xgboost.
Vertical Kmeans	vertical/kmeans	two-party or multi-party vertical kmeans
Vertical Sampler	vertical/sampler	two-party or multi-party vertical sampler