Machine Learning Prediction For The Safety Evaluation Of Adhesives In The Knee Meniscus Microenvironment

Medical tissue adhesives are often regarded as a new treatment method for meniscus tears.However,tissue adhesives are not easy to achieve clinical observation,and the development efficiency of new tissue adhesives can be greatly improved by using new three-dimensional tissue culture technology in vitro.Our aim is to generate a tissue adhesive safety evaluation data set based on the knee meniscus tissue culture technology,and apply it to a supervised machine learning classifier model to obtain the tissue adhesive safety evaluation results.In this paper,we first proposed a new kind of 3 d model of tissue culture,the new model adopts a multilayer structure,consider to embed the sensor structure model,in this model realizes the meniscus of cells and tissue adhesives drug molecules were culture experiment,based on the experimental design and build the organization adhesive for safety assessment of data sets,in the end,The machine learning binary classification algorithm is used to analyze and process the data set,and the value of the experimental data obtained from the in vitro experiment is mined out.In this study,tissue culture and new drug molecular research are both international cutting-edge technologies,which require high equipment and professional technology,and are difficult to implement.Also,open source data information is relatively small.The project uses the simulation data based on the knowledge graph technology to solve the problems of data set construction and insufficient data.The simulation data set with a capacity of 1600 pieces of data is used to conduct model training and evaluation for three binary classifier algorithms.Moreover,the stability of the classifier is observed by adjusting the data volume of the security category.The random forest classifier gets more excellent and stable results,with the accuracy rate of about 85%,and is not sensitive to the difference of data sets.The evaluation results show that the machine learning binary classifier can process the adhesive safety evaluation data set based on tissue culture well.The application scenarios designed by us are applicable to most drug-tissue culture fields,which can effectively reduce the number of biological experiments,the time and cost of new drug discovery,improve the efficiency of drug screening,and provide ideas and basis for computer processing of tissue culture experimental data sets,thus having practical application prospects and values.