An Efficient KNN Algorithm for the Mental Health Performance Assessment Using K-means Clustering

In today’s world, mental health assessment is very important and crucial due to the current lifestyle and other social factors. In this paper, an efficient K-nearest neighbor (KNN) approach was used for the mental health performance assessment using the k-means algorithm. Open-sourcing mental illness (OSMI) dataset was considered for experimentation from the Kaggle repository. The K value was computed randomly and compared with the error rates in KNN. Then the final K value has been selected which is based on the error rates. We have applied k-means algorithm before the assignment of top k rows. K-means algorithm helps it in the assignment of more appropriate class labels. Based on the distance measures, k-means are capable in the assignment of the data point in the best cluster group. Three different distance measures have been considered for the experimentation. The best value has been considered as the threshold for the class label assignment. The results depict the effectiveness of the combined approach. It shows the improvement in terms of K, epochs, and split ratios.