A study of momentum based on algorithms such as ARIMA_LSTM, HMM and Monte Carlo tests

In order to study the influence of "momentum" on daily life, this paper analyzes the data of a game. First, the data were cleaned up by quantitative letters. The dimensionality of the features is reduced by principal component analysis, and it is found that the server, the number of rallies and the speed of the car are the main influencing factors. Then, ARIMA-LSTM time series prediction analysis was used to eliminate the influence of short-term memory on the model and get the data of the main factors in the future. Logistic regression analysis was used with RMSE=0.47. A good fit shows that the higher the degree of influence of the server, the "momentum" increases. The Monte Carlo stochastic simulation experiment was used to determine the role of "momentum", and it was found that scores deviated too much from the Monte Carlo curve, with RMSE=3000.47, which was not in line with the Monte Carlo stochastic simulation experiment, and did not show a randomized distribution. Then, TOPSIS entropy weight synthesis evaluation method was used to quantify "momentum", and the three factors in the first question were used as the main evaluation indexes. Thirdly, the HMH Hidden Markov Chain is found to determine the probability of state continuation. Speed per hour, server, really counts as hidden set and "momentum" as found set. Summing the expected frequencies of each feature, the chi-square test shows p-value = 0, proving that "momentum" is not random, but there is a certain preference. Therefore, this study provides an effective basis for judging the development trend of some things in daily life, and can improve the development trend of things in people's daily life by influencing certain factors.