Enhancing The Accuracy In Predicting Wine Preference of Customers Using K-Nearest in Comparison With Random Forests
D. V. Sudheshna Sai, Terrance Frederick Fernandez · 2024
Improvement of the exactness in terms of when predicting wine preference by K-nearest (KNN ) than random forests (random forest) The code is written and run by using the dataset of the table format. Clinicalc.com is utilised to determine the test sample size which is estimated at a minimum of 2 samples per group or subgroup with a confidence interval of 85% with an alpha error rate of 0. accuracy of 05 and a G power of 85%. The findings reveal that the accuracy of the KNN of accuracy 85 87 percent yielded percent yielded the highest percentage. 00%. and random forest of accuracy of 78. 00%. The results of conducting the sample t-tests mean that, yes, there are differences among these groups, and the observed differences are statistically significant. In this case, the outcome indicates that the KNN algorithm has a better outcome than the random forest algorithm with statistically significant differences. We at last concluded that out of the available algorithm the KNN algorithm will be most suitable for us.