Design And Implementation Of Enterprise Growth Evaluation Model Management System Supporting Model Iterative Optimization

In the process of enterprise financing,service providers need to evaluate the growth of enterprises(stage of development,scientific and creative capabilities,etc.)to know their future development potential and competitiveness in order to match capital market products and services.The current method of evaluating the growth of enterprises is still based on due diligence and manual evaluation by experts in the field,which lacks effective support from the data level,system aspect and intelligent algorithm analysis.With the rapid development of machine learning and big data technology,machine learning technology can be used to improve the degree of automation in the process of evaluating the growth of enterprises,and provide more accurate data analysis and faster decision support for enterprises.Therefore,an enterprise has established a big data platform of science and technology enterprises,which contains machine learning evaluation models for enterprise growth stage,science and innovation capability,and financial risk.However,with the increasing demand of enterprise growth analysis,the system will integrate more machine learning models in the future,so there is an urgent need to establish a model management system that can support model iteration and optimization.This thesis designs and implements an enterprise growth evaluation model management system that supports iterative model optimization,through which users can manage machine learning models used in big data platforms,model training data,and server hardware resources used for model training.To solve the problem of machine learning model accuracy failure as business data grows,this paper proposes a data-driven iterative machine learning model-based approach,introducing domain experts to verify model results to form feedback,and performing model retraining and model version management based on data versions.Machine learning experts can use this data iteration method to complete the iteration of the training data-set,quickly train new enterprise growth evaluation models and deploy the models to the big data platform.Finally,this system enables the life-cycle management and iterative optimization of enterprise growth evaluation models.The system has been integrated into the big data platform of science and technology enterprises and put into practical use,and the system is in good running condition.By using this system,machine learning experts can quickly iterate and optimize various models of the original enterprise growth evaluation,which can effectively improve the accuracy and optimization efficiency of the application of enterprise growth evaluation models,and can reduce the difficulty of participation of domain experts,thus improving the accuracy and effectiveness of the application of models for enterprise growth evaluation.