[Retracted] Heterogeneous Network-Based Inductive Matrix Methods for Predicting Biomedical Gene Disease

Algorithm 1

Enhanced inductive matrix completion based on Katz.

Inputs: Gene and disease feature matrices X, Y, association matrix P, set of sampling subscripts Ω, gene similarity matrix G, disease similarity matrix D, parameters β, δ, ρ, λ, and the number of iterations Max_iter

Output: Predicted correlation matrix + XZ

1. Calculate (C) according to equation (5)

2. Calculate the residual matrix R

3. Initialize Z₀ = 0

4. For k= 0 to Max_iter

5. Update Z according to equation (17)

6. End for

7. Return (C) + XZ