Task: Email Spam Rating
Dataset: spambase.data from the UCI Machine Learning Repository
Algorithms used:
- k-nearest neighbors (kNN)
- Artificial Neural Networks (RNA)
- Naive Bayes (NB)
- Linear discriminant analysis (LDA)
Conclusion: Among the four classifiers, the RNA was the best one, with an accuracy of 0.925, in contrast, the classifier that least adapted to the data was the NB with accuracy of 0.686.
Task: Mushroom Rating
Dataset: mushroom.data from the UCI Machine Learning Repository
Algorithms used:
- Decision tree (AD)
- Random Forests (RF)
- Support Vector Machines (SVM)
- Logistic Regression (RL)
Conclusion: Among the four classifiers, the RF and SVM were the best, with precision of 1.000, in contrast, the classifier that had the least precision was the AD with the value of 0.989.