USING A LOGICAL MODEL TO PREDICT THE FUNCTION OF GENE: A SYSTEMATIC REVIEW

Mining large data sets using machine learning approaches often results in models that are difficult to interpret and not amenable to generating hypotheses that can be tested experimentally. Scientists attempt to study how complex systems behave as a whole rather than focusing on isolated components. It also gave researchers the ability to simultaneously monitor interactions between individual components of the experimental data. In biology Gene expression data can also be used to model a transcriptional network that is characterized by interactions between thousands of genes and proteins. In this study, the most important machine learning methods used in data mining and analysis to predict the species and new forms of the genes through multiple machine learning methods were discusses. Gene activities were studied to know how they evolve. It is suggested that the first steps in modern machine learning approach is the supervised learning process, and then focused on continuous learning per practical interactions and outcomes prediction. By combining seven word similarities and five knowledge-based similarities, MFR worked. Based on previous experience, he kept gene pairs with high similarity of expression and similarity and proposed them by models. Initially, a 10-fold cross-validation of the MFR model was carried out. Then the MFR model was used on the Gene Friends and DIP datasets for a validity test and other validations. Finally, the MFR model is used to create cancer gene networks and to forecast gene activity