| In recent years,as the economy and society continue to develop,small and medium-sized enterprises(SMEs)in China have increasingly contributed to the national economy.In 2021,SMEs in China contributed over 60% of the GDP,70% of patents,and 80% of job positions.This demonstrates the close relationship between the growth of SMEs and the development of the national economy.Furthermore,due to the varying levels of growth among SMEs,there is an urgent need to evaluate and distinguish their growth levels.In the field of growth evaluation research,current scholars mostly focus on constructing comprehensive indicators or using factor analysis methods for evaluation.However,these methods may not fully consider the subjective consciousness and behavior of enterprise managers,and the evaluation level is relatively one-dimensional,lacking comprehensive consideration of enterprise operation and management processes.Moreover,there is a lack of quantitative research on the evaluation of corporate growth at present.Therefore,this article proposes a scoring card construction method based on machine learning algorithms to improve the evaluation methods for the growth of SMEs.This literature review examines the current state of research on growth influencing factors and growth evaluation methods using a qualitative research approach.The author identifies influencing factors and measurement indicators of enterprise growth from five dimensions:internal governance,external governance,industry status,business status,and policy environment.The study uses data from listed companies on the Small and Medium-sized Board and Chi Next Board as the research sample,and fills in missing values using the interpolation method and random forest estimation method.Feature variables in the data are screened using WOE boxing,IV value comparison,and correlation analysis methods.Exploratory data analysis is conducted using quantitative analysis methods such as ratio analysis and trend analysis.After testing all machine learning models,the author selects logical regression,random forest,XGboost,and GBDT algorithms to build growth evaluation models.Results show that logical regression models and random forest models have the best fitting effect.Finally,the author uses the scoring card method to build a quantitative scoring system for SME growth based on the logical regression and random forest algorithms.After conducting the research,the following conclusions are drawn:(1)The method of constructing a scoring card to evaluate the growth of small and medium-sized enterprises,adopted in this article,is both feasible and effective.The logical regression score card is more logical than the random forest score card,while the random forest score card has a better differentiation ability.(2)The growth rating card for small and medium-sized enterprises can provide rating results and targeted adjustment suggestions for enterprises.Moreover,the rating results and analysis of key factors can enable the government to develop more effective support policies.(3)Compared to previous evaluation methods such as comprehensive indicator evaluation and principal component analysis,the evaluation method constructed in this article takes into account more comprehensive factors,which enhances its practicality and operability.Through the use of machine learning algorithms and scoring cards,this article constructs a new evaluation system for small and medium-sized enterprises,enriching and improving research on enterprise growth evaluation.Furthermore,it proposes and confirms the feasibility and effectiveness of quantitative scoring research methods,which contributes to the sustainable development of small and medium-sized enterprises in China.The shortcomings of this article lie in the need to further improve the selection of growth influencing factors and growth measurement methods,as well as the scope of sample selection for small and medium-sized enterprises.In recent years,the evaluation system of this article has not included the impact factors of the epidemic. |
