/Data_Science_MOOC_Exercise

My programming solutions for differents interesting data science exercises

Primary LanguageMATLAB

# Data_Science_MOOC_Exercise Exercise from "Introduction to Data science" (Coursera), "Fondamentaux du Big Data" (Fun mooc), "Machine Learning" (Coursera, Andrew Ng).

0) Twitter sentiment analysis (Python)

Challenge: Give a positive or negative sentiment value for tweets
Techniques: Data wrangling, python data manipulation, use of tweeter API and data
How to Run:

1) Linear Regression (Octave)

Challenge: Guess the profits for a food trucks
Techniques: gradient descent, data plot, cost vizualization, feature normalization
How to Run: type "ex1" in Octave shell in the right repository (need to download Octave: https://www.gnu.org/software/octave/download.html)

2) Logistic Regression (Octave)

Challenge: Guess the chance of admission in university for applicants.
Techniques: data vizualization, sigmoid, linear logistic regression, feature mapping, polynomial logistic regression
How to Run: type "ex2" in Octave shell in the right repository.

3) Multi-class classification and Neural Network (Octave)

Challenge: Recognize handwritten digit
Techniques: All-vs-one classification, All-vs-one prediction, neural network, feedforward propagation
How to Run: type "ex3" in Octave shell in the right repository, and also type "ex3_nn" for part 2

4) Neural Network Learning (Octave)

Challenge: Recognize handwritten digit
Techniques: regularized neural network, feedforward propagation, backpropagation, gradient checking
How to Run: type "ex4" in Octave shell in the right repository

5) Regularized Linear Regression and Bias vs Variance (Octave)

Challenge: Guess how much water flowing out of a dam and debug
Techniques: regularized linear regression, learning curves, polynomial regression, cross validation for regularization parameters
How to Run: type "ex5" in Octave shell in the right repository

6) Support Vector Machine (SVM) (Octave)

Challenge: Guess if a mail is a spam or not
Techniques: SVM with gaussian kernel, feature extraction
How to Run: type "ex6" in Octave shell in the right repository and also type "ex6_spam" for part 2

7) K-means Clustering an Principal Component Analysis (PCA) (Unsupervized Learning) (Octave)

Challenge: Image compression
Techniques: K-means, PCA
How to Run: type "ex7" in Octave shell in the right repository and also type "ex7_pca" for part 2

8) Anomaly Dectection and Recommender Systems (Octave)

Challenge: Detect anomalous behavior in server computers. And build a movie recommender.
Techniques: gaussian distribution, collaborative filtering learning algorithm, learning movie recommendations
How to Run: type "ex8" in Octave shell in the right repository and also type "ex8_cofi" for part 2