Research And Implementation Of Noncoding RNA Association System Based On Transformation Hybrid Recommendation Method

In recent years,the popularization and development of artificial intelligence has been like a speeding train,and more and more researchers who study the relationship between biomacromolecules are also taking the train.Because compared with traditional biomedical experiments,it can greatly save the research cost and energy.Recently,non-coding RNA has been found to play a crucial role in development,metabolism,disease and other life activities.Non-coding RNAs are RNA that do not code for proteins and vary in length from dozens to hundreds of nucleotides.Both long non-coding RNA and micro RNA belong to non-coding RNA,and the long non-coding RNA can regulate micro RNA,so as to have a certain impact on diseases.Therefore,the use of recommendation algorithm to study the association between long non-coding RNA and micro RNA has become a hot topic.In this experiment,the transformation hybrid recommendation method is used.The algorithm first uses the Gaussian algorithm and the Needleman-Wunsch algorithm to calculate the similarity of non-coding RNA,and uses the transform coefficient ? to calculate the similarity K nearest neighbor matrix decomposition,confidence algorithm and cosine similarity algorithm combination.According to the change of transformation coefficient ?,different algorithms are used.Then,this experiment applied the transformation hybrid recommendation method in the field of bioinformatics,recommending the association between TOP long non-coding RNA and micro RNA.The AUC values of 0.8217,0.8810 and 0.9814 were obtained in the interval of their respective transformation coefficients.Furthermore,this experiment tests the applicability of the transformation hybrid recommendation method based on the test sets of different non-coding RNA types,and makes case analysis of the three experiments to verify the effectiveness of the method.In summary,the results show that the method will reduce the computational efficiency of recommendation compared with a single recommendation algorithm,but it is superior to other single recommendation algorithms in accuracy,which shows the good applicability and recommendation performance of the transformation hybrid recommendation method.Focusing on functional and non-functional requirements,this system divides the non-coding RNA correlation recommendation system into registration and login module,data preparation module,algorithm recommendation module,case analysis module,case recommendation module and system management module.In order to make the front page simple and beautiful,this system USES Bootstrap and Sweetalert framework,and uses Python and PHP language to build the background.After comprehensive testing,the entire system was deployed to the aliyun site to ensure the long-term operation of the system.