Scribd
Upload
3 views1 page

ML Classification Notes Student Distribution

The document outlines key concepts in machine learning classification, including performance metrics like accuracy, precision, and recall, as well as the use of confusion matrices and ROC curves. It discusses ensemble methods such as bagging and boosting, multilayer feedforward neural networks, support vector machines, and the k-nearest neighbor algorithm. Each method is briefly explained along with its functionality and common techniques used.

Uploaded by

Swathi Reddy
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
3 views1 page

ML Classification Notes Student Distribution

The document outlines key concepts in machine learning classification, including performance metrics like accuracy, precision, and recall, as well as the use of confusion matrices and ROC curves. It discusses ensemble methods such as bagging and boosting, multilayer feedforward neural networks, support vector machines, and the k-nearest neighbor algorithm. Each method is briefly explained along with its functionality and common techniques used.

Uploaded by

Swathi Reddy
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

Machine Learning Classification Notes

1. Metrics for Evaluating Classifier Performance


Classifier performance measures how well a model predicts classes. A confusion matrix contains
TP, TN, FP, FN. Accuracy=(TP+TN)/Total. Precision=TP/(TP+FP). Recall=TP/(TP+FN).
Specificity=TN/(TN+FP). F1-score balances precision and recall. ROC curve plots True Positive
Rate against False Positive Rate. AUC closer to 1 indicates a better classifier.

2. Ensemble Methods
Ensemble learning combines multiple models to improve accuracy. Bagging trains models in
parallel (e.g., Random Forest). Boosting trains sequentially to correct errors (AdaBoost, Gradient
Boosting). Voting combines predictions, while Stacking uses a meta-model.

3. Multilayer FeedForward Neural Network


An MLFFNN contains input, hidden and output layers. Data moves only forward. During training,
forward propagation computes outputs and backpropagation updates weights to reduce error.
Common activation functions are ReLU, Sigmoid and Tanh.

4. Support Vector Machines


SVM finds the optimal hyperplane with maximum margin between classes. Support vectors are the
nearest points to the boundary. Kernels such as Linear, Polynomial and RBF allow nonlinear
classification.

5. k-Nearest Neighbor
KNN is a lazy learning algorithm. To classify a new sample, compute distances to training samples,
choose K nearest neighbors, and assign the majority class. Common distance measures are
Euclidean and Manhattan.

You might also like