Algorithm 1.

Algorithm to extract the sample representation matrix for each data type.

Input: the profile of i th data type, i.e. $X^i=\left[x_1^i,x_2^i,\mathrm{...},x_n^i\right]$, tuning parameter λ, and nearset neighbors parameter K. Output: the sample representation matrix W i of i th data type. 1. Obtain neighbors in data X i using K-nearset neighbour method, and assign the parameter Ω 2. Solve the equation (1) by updating (4), (5), (9)-(12) until the iteration converges and obtain the optimal Z* 3: Construct the sample similarity matrix W I by $W^i=\left(\left|Z^{*}\right|+\left|Z^{*T}\right|\right)/2$