Code & Results

The XGBoost results offer valuable insights into the performance of the model for anxiety severity classification. After applying hyperparameter tuning, XGBoost achieved higher accuracy compared to models like SVM, with a noticeable improvement in distinguishing between low and high severity classes. XGBoost is a powerful ensemble method that leverages decision trees, and its ability to handle non-linear relationships and complex patterns within the data contributed to its success. The model’s performance indicates that the features in the anxiety dataset contain enough information for XGBoost to exploit, yielding a more accurate classification. The confusion matrix and feature importance analysis provided further insights, revealing which factors (such as stress levels, caffeine intake, and sleep patterns) played a significant role in predicting anxiety severity. These findings suggest that XGBoost is a more suitable model for this type of task, as it can capture intricate patterns in the data that other models, like SVM, struggled to identify.