Research and Application Based on Time Series Analysis with Least Squares Regression

In this paper, a simple model in the exponential smoothing model is developed based on the time series of the given data, which has a good fit with an R
²
of 0.984 and passes the residual test. Finally, the number of outcomes reported on March 1, 2023 is expected to be within the 8% confidence interval of (7,478,30763). The specific number of reports was approximately 19,120, and a Spearman correlation analysis was subsequently performed to plot a heat map of the correlation coefficient. The analysis revealed no correlation between the two. Then, ten randomly selected words of data were retained for reliability calculations, and then seven expressions were obtained with six attributes as independent variables and correlation percentages (1, 2, 3, 4, 5, 6, X) of partial least squares regression analysis were performed daily as dependent variables. Then the eerii6 attributes were extracted for expressions and rounded to obtain 1: 1%, 2: 8%, 3: 23%, 4: 28%, 5: 25%, 6: 12 %, and X: 2%. Finally, the retained data are put into the expressions and their reliability is calculated. The relative error between the predicted data and the actual data was 7.53%, which means the reliability was 92.47%.