# MMGINCDA

**Repository Path**: Tomhappy/MMGINCDA

## Basic Information

- **Project Name**: MMGINCDA
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: main
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2026-04-11
- **Last Updated**: 2026-04-11

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# MMGINCDA

We propose a new computational model, Multiple similarity and multiple kernel fusion based on graph inference network for predicting circRNA-disease associations. This is the implementation of MMGINCDA:



# Environment Requirement

    python == 3.9.13
    torch == 2.5.1
    torch-genometric == 1.4.2
    matplotlib == 3.5.2
    networkx == 2.8.4
    numpy == 1.21.6
    pandas == 1.4.2
    scipy == 1.9.1
    
# Dataset

We performed 5-fold cross validation on four datasets. Dataset1-4 are from CircR2Disease database, CircRNADisease database, Circ2Disease database, and CircR2Disease v2.0, respectively. We divided the known circRNA-disease associations into five equal parts and stored them in .txt files.

# Model

    MMGINCDA.py: This file contains the main function. The paramaters of MMGINCDA are also adjusted in this file.
    GKS.py: This file contains calculating the Gaussian kernel similarity.
    Global_similarity.py: This file contains calculating the global similarity.
    LKS1.py: This file contains calculating the Laplace kernel similarity.
    Local_similarity.py: This file contains calculating the local similarity.
    SKF1.py: This file records the process of model fusion.
    known.py: This file contains the knonw and unkonw circRNA-disease associations.