Dynamic Forecasting Analysis of College Students’ Employment and Entrepreneurship Market Trends Based on Long and Short-Term Memory Network Models

This paper first establishes a hierarchical employment and entrepreneurship quality coordinated development analysis system, through the use of time series ARIMA model and neural network LSTM model, respectively, for short-term prediction of college students' employment and entrepreneurship market dynamics trends.In view of the prediction defects of a single model, this paper adopts the optimal weighted combination prediction method to optimize the combination of two single models, obtains the ARIMA-LSTM combination measurement model, and analyzes the application effect of the three models with actual cases.The results show that the relative errors of single models ARIMA and LSTM in predicting the employment and entrepreneurship of college students in the validation set are 9.52% and 10.13% respectively, and the R is 83.0378 and 81.2749, which shows that the predictive effect of the models is general.The ARIMA-LSTM combination model is significantly better than the single model ARIMA and LSTM in terms of the accuracy and stability of the prediction of college students' employment and entrepreneurship, at this time, the correlation indexes of the three models of ARIMA, LSTM and ARIMA-LSTM are 0.8064, 0.7959 and 0.9773, respectively, which can be seen that the combination model can effectively integrate the two single model s linear forecasting ability and nonlinear time series modeling advantages, thus improving the accuracy and reliability of the ARIMA-LSTM model in predicting the dynamics of college students' employment and entrepreneurship market trends.