Comparison of Prediction Accuracy of Logistic Regression and Random Forest in Weather Data Time Series Analysis

The purpose of this study is to compare the performance of logistic regression model and random forest model in meteorological time series data analysis and analyze the difference when applying them to precipitation probability prediction. Using logistic linear regression model and random forest algorithm, we examined how major weather variables (Temperature, Dew Point, Humidity, etc.) contribute to precipitation occurrence prediction. In the experiment, the logistic regression accurately predicted 5,545 out of 6,812 samples, achieving a prediction accuracy of 81.4%. In contrast, the random forest model, an ensemble technique that can learn nonlinear interactions, utilizes an ensemble learning technique that combines multiple decision trees to improve prediction performance. In the experiment, the random forest model achieved a prediction accuracy of 90.61% on the test data, showing about 9% higher performance than the logistic regression model. In summary, in time series data analysis, it is important to select an appropriate model considering the linearity and complexity of the data. The logistic regression model is suitable for situations that require a simple structure and quick interpretation, while the random forest model excels in problems that include nonlinear data and complex interactions. Keywords Logistic regression, Time series data, Random forest, Weather prediction and Machine learning