Rating Analysis and Prediction Based on K-Nearest Neighbor Regression

With the continuous development and progress of society, television has entered thousands of households early. In order to better analyze TV ratings, this paper proposes a rating analysis method based on K-nearest neighbor regression for the analysis and prediction of TV ratings. The method first captures the historical ratings data published on the Internet; then analyzes the original data, and visually displays the relationship between the features to be analyzed; then selects the K-nearest neighbor regression algorithm as the main basis for analysis, respectively Real-time audience rating data), audienceRatingAvg (average audience rating data), audienceRatingMax (highest audience rating data), and audienceRatingMin (minimum audience rating data) modeling to obtain the best fitting regression curve; The value of the ratings data. After many tests and analysis, the model can predict the viewership data very well, and can provide fast analysis and processing methods for those who need it.