Abstract
Wordle is a popular puzzle that The New York Times currently provides every day, and it has a high popularity. Among them, the number of results reported every day, the characteristics of words and other data have attracted widespread attention. This paper first used the ARIMA model to predict the number of daily reported outcomes and found that it was only accurate for the linear part of the data. Then, this paper used the LSTM neural network model to predict, and found that the LSTM model can predict the nonlinear part of the data well, which just makes up for the deficiency of the ARIMA model, and the predicted results are basically consistent with the original data. The data range of March 1st is [17586.36, 44379.83]. Further, this paper adopted the LSTM neural network model based on genetic algorithm optimization, which can solve the over-fitting problem that may occur in the LSTM neural network due to too few data sets. Finally, the SVM multi-classification model are used. According to the quantified word feature labels, the difficulty of words is divided into three categories: hard, medium, and easy. Using existing data tests, it’s proved that the classification accuracy is very high.
Publisher
Darcy & Roy Press Co. Ltd.
Reference10 articles.
1. Liu Chengliang. Research on Air Quality Index Evolution Prediction Model Combining GCN and LSTM [D]. Nanjing: Nanjing University of Posts and Telecommunications. 2022.
2. Longfuhai. Study on the feature selection method based on the optimization of genetic algorithms [D]. Guiyang: Guizhou National University, 2022.
3. Okkalioglu Murat. Imbalance text classification with relative imbalance ratio [J]. Expert Systems with Applications,2023, Volume 217, Issue.
4. Luo Mao. Research on Support Vector Machine Optimization Algorithm Based on Improved Multiverse Algorithm [D]. Changchun: Jilin University, 2022.
5. Hans van Halteren. Improving Accuracy in Word Class Tagging through the Combination of Machine Learning Systems [J]. Computational Linguistics ,2001, 27 (2): 199–229.