Vec2image: an explainable artificial intelligence model for feature representation and classification of high-dimensional biological data by vector-to-image convention
Vec2image: Vec2image for MATLAB is a free, open-source code for classification and feature selection
(1) Two input files are required: 1) a features-by-samples data table, 2) a list of feature labels. (see Example dataset1 and dataset2).
(2) Main.m is the main program to execute Vec2image algorithm. You can execute the program step by step
(3) Output: The output file res.mat contains the well-trained model in out.mat, the predicted accuracy of test data and the importance/impact score of each feature.
(4) dataset1 and dataset2 is provide to verify the code.
Note :
(1) We will use the trainNetwork() function in Matlab when we train a CNN model, please note your matlab version (>= MTALAB2020a).
(2) You can train on either a CPU or a GPU (makeObjFcn5.m at Line 131). For image classification, you can train a single network in parallel using multiple GPUs or a local or remote parallel pool to save training time. To use a GPU for deep learning, you must also have a supported GPU device.
(3) The MaxEpochs is set to 100 in makeObjFcn5.m at Line 130. Please change this as desired. This will affect the processing time of training.
Now, the package code V0.2 is freely accessed.
Contact Email: tanghui2020@scut.edu.cn; zengtao@sibs.ac.cn
##############################################################################################################################################################
-
V0.2 Supporting umap now, in addition to kPCA, tSNE
Supplying the python implementation of ResNet for Vec2image, which is based on PyTorch
##############################################################################################################################################################
Reference: Tang H, Yu X, Liu R, Zeng T. Vec2image: an explainable artificial intelligence model for the feature representation and classification of high-dimensional biological data by vector-to-image conversion. Brief Bioinform. 2022 Jan 31:bbab584. doi: 10.1093/bib/bbab584. PMID: 35106553