Multi-view multi-classification method for subcellular protein localization
A protein and subcellular technology, applied in the field of multi-view and multi-classification of subcellular protein localization, can solve the problems of unbalanced data quantity difference, poor effect, and inability to make full use of multi-view data learning, etc., to achieve high accuracy and precise prediction Effect
Pending Publication Date: 2021-02-26
GUANGDONG UNIV OF TECH
View PDF0 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0005] However, from the perspective of machine learning, the protein localization data set has several characteristics: the data set has too many classes, belongs to the "multi-label" data set, and is too "unbalanced" (the number of data in each class varies greatly ), for a single-view method such as SVDD, it cannot make full use of multi-view data for learning, and the final effect is not as good as the multi-view method
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0053] The present invention will be further described below in conjunction with specific embodiment:
[0054] A multi-view and multi-classification method for subcellular protein localization described in the embodiment of the present invention firstly trains and optimizes the learning classifier model, and then predicts the position of the protein in the cell through the optimized learning classifier model; while training When optimizing the learning classifier model, multi-view analysis is performed on the training data of subcellular protein localization, and the training data containing multiple views are mapped to the same feature space for SVDD calculation. The multi-view problem is transformed into a single-view problem.
[0055] The specific steps are as follows:
[0056] S1. Training and optimizing the learning classifier model. In the subcellular protein localization training data set, divide the data set into a target data set and a negative data set, and find the...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses a multi-view multi-classification method for subcellular protein localization, and the method comprises the steps: firstly training and optimizing a learning classifier model,and then predicting the position of protein in a cell through the optimized learning classifier model; when the optimization learning classifier model is trained, carrying out multi-view analysis on training data of subcellular protein localization, mapping the training data containing multiple views into the same feature space for SVDD calculation, in the calculation process, adding weights to the views, and converting the multi-view problem into a single-view problem through the weights to be solved. Compared with a single-view-angle method, the method has higher accuracy.
Description
technical field [0001] The invention relates to the technical field of subcellular protein localization, in particular to a multi-view and multi-classification method for subcellular protein localization. Background technique [0002] Subcellular protein localization, the position of a protein within a cell, is one of the key functional characteristics of a protein. Due to the need for large-scale genomic analysis, an automatic and efficient method for the prediction of protein subcellular localization is required. [0003] Outlier detection, also known as outlier detection, assumes that all training data comes from the same class. Therefore, the anomaly detection algorithm obtains a compact bound for the target data. A training set of objects and detection of (new) objects like this data description can be used for outlier detection to detect unusual objects from a dataset. Often these outlier data show abnormally large or small feature values compared to other trainin...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/62G06N3/08G16B15/00G16B40/00
CPCG06N3/08G16B40/00G16B15/00G06F18/2431G06F18/24G06F18/214
Inventor 钟光正肖燕珊刘波
Owner GUANGDONG UNIV OF TECH