Prediction Of Relationship Between Non-coding RNA And Disease Based On Association Rule Mining

With the development of functional research of non-coding RNA in biology,more and more evidences shows that non-coding RNA plays an important role in the prevention,occurrence and development of diseases,especially in the proliferation and invasion of malignant tumors.Functional prediction based on large amounts of gene regulatory network data has become a hot topic in industry and academia,especially in predicting nc RNA,which is closely related to the occurrence and development of diseases.This study is of great biomedical value in identifying candidate nc RNAs for disease diagnosis,treatment and prognosis.However,with the increasing scale and complexity of these interconnected networks,it is extremely challenging to accurately locate nc RNA-diseases with significant association.Statistical and machine learning methods have become a powerful tool for solving such problems because of their high efficiency and stability.Based on gene regulation network data,this study introduces statistical methods and machine learning techniques,including supergeometric distribution method,Gaussian kernel function and improved particle swarm optimization algorithm,and proposes a new intelligent prediction model.On this basis,the study through related research,experiments and analysis of the relationship between non-coding RNA and disease,achieved a significant nc RNA-disease relationship prediction model and carried out relevant experiments to verify,so this study has important academic significance.The work of this paper is described as follows:(1)Research on predictive models of significant nc RNA-disease relationships in gene regulatory networksBased on the characteristics and research strategies of gene regulation network data,the existing prediction models are introduced,analyzed and compared,and the functional similarity is defined by the data analysis angle.The research ideas and overall strategies of the prediction model are presented progressively.(2)Construct an efficient and accurate nc RNA-disease relationship prediction modelA hybrid prediction model based on statistical method and improved particle swarm optimization algorithm is proposed.Firstly,hypergeometric distribution and Bonferroni method are used to determine the relationship between candidate non-coding RNA and disease.Secondly,similarity matrix and distance matrix are constructed by using Gaussian kernel function and distance criterion.On this basis,non-coding RNA functional modules are clustered based on improved particle swarm optimization algorithm.Finally,the disease sets associated with non-coding RNAs in each functional module are intersected to precisely lock the relationship between each non-coding RNA and disease.(3)Model development,verification and evaluationAccording to different data sources and different types of non-coding RNA,the data were divided into three different experimental data groups.The optimal setting parameters of the model are obtained through experiments,and the experimental results are analyzed and evaluated.The experimental results show that the model can accurately predict the relationship between non-coding RNA and disease through gene regulatory network,and achieve good results.