This programming exercise was done as part of Coursera's Machine Learning Course (Stanford University), taught by Prof. Andrew Ng.
- Built a classification model that estimates an applicant's chances of admission into a university based on exam results using logistic regression
- Historical data from previous applicants was used as the training set
- Used Octave's fminunc optimization solver to find optimal parameters of the model
- Computed training accuracy of the classifier
Predict quality assurance result (passed/not) with test results for microchips from a fabrication plant
- Implemented regularized logistic regression to predict whether microchips from a fabrication plant pass QA based on test results
- Used a dataset of test results on past microchips to build the model
- Performed feature mapping to obtain a more expressive classifier and implemented regularization to combat overfitting
- Used Octave's fminunc solver to learn the optimal parameters
- Studied the effect of the regularization parameter on the fit