Python Implementation of Logistic Regression and Naive Bayes Algorithms -- 2
- 状态 关闭的
- 预算 $10 - $30 USD
- 总竞标 1
(a) Logistic Regression. Report the accuracy on the test set. Vary the learning rate (η) and report the results for 3 different learning rates. Report the confusion matrix on the test set. Since this requires a k-class logistic regression, predict if the class is 1 or not.; to achieve this, create a new version of the data set where for all the examples where the class label is not 1, you assign a new class label (say 0). Thus now your binary task is predicting whether class 1 is true or not. Do the same for the test set as well.
(b) The counting based Naive Bayes classifier: Assume Laplacian correction. Again, treat the task as binary and report the results as a confusion matrix.
Data: from the UCI Zoo data set (zoo-train and zoo-test). There are 16 features (the first 16 columns) and the class labels are in the last column. There are 7 classes (numerically specified as class 1 to 7). All features are binary except for feature 13, which is a categorical variable with possible values 0,2,4,5,6,8. Note that to create binary split, please use the one-vs-rest approach.
- The New York Times
- Wall Street Journal
- Times Online