STRESS PREDICTION USING ENHANCED FEATURE SELECTION AND KNN MODEL

The necessity of becoming aware of one's mental and physical well-being has been growing rapidly in light of the modern economy's enhanced conditions. Numerous significant factors contribute to mental health problems like depression and anxiety. Medical-based systems like an electroencephalogram, an electrocardiogram, and a heart rate variability analysis were utilized in the previous models. The analysis of a person's brain activities and mental state will be done by taking into account the corresponding readings and recordings of each machine used here. The results of specific approaches are gathered, which suggest different brain signal variances and nerve disorders. The implementation of this paradigm is more difficult, time-consuming, and expensive when compared to other models. The KNN (K-Nearest Neighbor) machine learning model is proposed here with an enhanced feature selection, a non-parametric technique that employs similarity to categorize or determine how each unique data point will be grouped. The enhanced feature selection model decreases the input variable and ignores the noise data. Filtering techniques like forward filtering and backward filtering are applied after preprocessing techniques.Here KNN predicts the outcome with an accuracy of 99% and is considered the best classifier.