Algorithm 1. Removing all dependent and redundant features

Input: Data $X\in R^{n\times d}$, where $n$ and $d$ are samples and features, respectively. 1-Rank the features based on information gain: $\{F_1,F_2,\dots ,F_d\}$, where $F$ represents a feature and $F_1>F_2>\dots >F_d$ in terms of information gain. 2-Select the first feature as selected feature: $F_1$, $X^{\prime }=\left\{F_1\right\}$ 3-for  $i=2:d$, #For all features $\left\{F_2,\dots ,F_d\right\}$ 4-Obtain rank $X^{\prime }={\{F}_1,F_i\}$ 5-if it is full rank 6- $X^{\prime }\leftarrow X^{\prime }\cup \left\{F_i\right\}$: keep $F_i$ 7-end if 8-end for Output: Selected features, $X^{\prime }\in R^{n\times d^{\prime }}$, where $d^{\prime }$ shows the number of selected features after removing redundant features and $d^{\prime }<d$.