Comparison of Support Vector Machine and K-Nearest Neighbour Algorithm for Accurate text Classification
A.V.Naresh Babu T L V Naga Lathish, T Devi. · 2022 4th International Conference on Inventive Research in Computing Applications (ICIRCA) · 2022
The proposed work focuses on performing text classification analysis using support vector machine (svm) as well as k-nearest neighbour algorithm (knn). Materials and Methods: Accuracy is analysed for emotion dataset with count of sentences 519. Classification of emotions is done using support vector machine with size$\boldsymbol{(\mathrm{n}=32)}$and k-nearest neighbour algorithm (knn) with size$(\mathbf{n=32})$, obtained using the g-power value 80%. Results: svm accuracy is 94.29% which is comparatively higher than knn with accuracy of 90.80%. the significant value of accuracy is 0.999$\boldsymbol{(\mathrm{p} > 0.05)}$and loss is 0.001$\boldsymbol{(\mathrm{p} < 0.05)}$. Conclusion: The proposed model indicates that accuracy obtained from model using SVM is higher than k-nearest neighbour algorithm.