Journal Articles (All Issues)

Abstract

The ultimate goal of this paper is to make the logistic regression model learn train data so that it can predict our subjects’ vote in the last presidential election. When it comes to categorical data such as gender and regional places, it is necessary to readjust them. By using one hot encoding, we readjusted them. A point to note is that cross validation without random is better than cross validation with random. More specifically, the logistic regression model works well with cross validation without random, whereas it does not work well with cross validation with random. A further point to note is that when C (a parameter) was 100, the logistic regression model obtained the best score (100%) in train data and test data by using grid search. On the other hand, when C was 23, the logistic regression model obtained the best score (100%) in train data and test data by using random search. A major point of this paper is that when C is 10, the accuracy rate of the model logistic regression is 100% in both of train data and test data. Finally, the Receiver Operating Characteristic (ROC) analysis tells us that the space between true positive and false positive is big. This in turn implies that the accuracy rate of the classification model logistic regression is high. It is clear from our findings that the logistic regression model is good enough and that it works well for train data and test data.

References

[1] Kang, N. (2023a). K-Pop in BBC News: A Big Data Analysis. Advances in Social Sciences Research Journal 10(2), 156-169. [2] Kang, N. (2023b). K-Dramas in Google: A NetMiner Analysis. Transaction on Engineering and Computing Sciences 11(1), 193-216. [3] Kang, N. (2023c). A Comparative Analysis of Tolerate and Put up with in the COCA. Semiconductor and optoelectronics 42(1): 1468-1476. [4] Kang, N. (2023d). Sure of and Sure about in Corpora and ChatGPT. Journal of Harbin Engineering University 44(7): 1347-1351. [5] Kang, N. (2023e). Turn out adj and Turn out to be adj in the Now Corpus and ChatGPT. Journal of Harbin Engineering University 44(8): 825-831. [6] Kang, N. (2023f). Care for and Like in Corpora and ChatGPT. Semiconductor and optoelectronics 42(2): 188-198.