About

Udacity Nanodegree Machine Learning Engineer assignments & note-taking.

Notes

1. Decision Boundaries : Plot decision boundaries on different classifiers.
2. K Cross Validation : Use KFold to split data into train-test sets for cross validation.
3. Grid Search : Use grid search to tune hyper-parameters of decision tree classification.
4. Linear Regression : Predict life expectancy based on BMI.
5. Multiple Linear Regression : Predict Boston house-price.
6. Perceptron Algorithm : Apply perceptron algorithm to separate classified dataset on optimizing weight & bias.
7. Decision tree : Titanic Survival Exploration with Decision Trees.
8. Naives Bayes : Build a SMS Spam classifier by using bag of words.
9. Naives Bayes 2 : Classify the emails as written by which author based only on email content.
10. Support Vector Machine : Simple example of SVM classification with kernel rbf.
11. Ensemble methods : Use AdaBoost with base of decision tree to classify terrain data.
12. Gradient Descent : Use gradient descent algorithm to find the boundary of dataset.
13. Polynomial Regression : Predict laptop battery life with polynomial regression with different degrees.
14. Logistic Regression : Predict gaussian dataset using logistics regression and visualize the computed model.
15. Neural Network MLP : Predict student admissions to graduate school in Keras.
16. Neural Network MLP 2 : Classify digit images (0 to 9) from the MNIST database in Keras.
17. Neural Network MLP 3 : Analyze IMDB movie data to predict sentiment analysis in Keras.
18. Clustering : Use k means clustering to explore Netfix data from movie ratings.
19. Clustering : Conduct hierarchical clustering on the Iris dataset.

Assignments (Project to be submitted for review)

1. Titanic Survival Exploration : Find decision tree rules based on visual exploration (barchart, histogram).
2. Predicting Boston Housing Prices : Evaluate predictive model for Boston Housing Prices.
3. Finding Donors for CharityML : Use several supervised algorithms to predict individual that makes > $50,000.
4. Dog Breed Recognition : Predict dog breed using deep learning CNN.