Page 12 - GPD-2-1

Gene & Protein in Disease AI-based drug repositioning

understand. Matrix completion is the decomposition of Xuan et al. proposed a prediction method for disease-
an incomplete matrix to obtain two or more submatrices, drug associations, and this method uses a non-negative
which are multiplied together to obtain a new matrix matrix decomposition (DivePred) technique. To indicate
that approximates the original missing matrix, and then, the characteristics of drugs from various perspectives,
the values in the new matrix are used to fill the missing DivePred incorporates the similarities of various drugs .
[34]
values in the missing matrix. Many matrix-based drug A low-rank matrix complementation method was proposed
repositioning methods have been proposed by researchers by Yang et al. This method can use the regularization term
in recent years, as shown in Table 3. to reduce the similarity noise, while the resultant values

A B

C E

D

Figure 4. (A-E) Overall workflow of multiview learning with matrix completion.

Volume 2 Issue 1 (2023) 6 https://doi.org/10.36922/gpd.v1i3.201