Machine Learning- Coursera - Stanford - Programming Exercises

IV. Linear Regression with Multiple Variables (Week 2) Programming Exercise 1

1) Warm up exercise  [ warmUpExercise.m ]

 2) Computing Cost (for one variable) [ computeCost.m ]

 3) Gradient Descent (for one variable) [ gradientDescent.m ]

 4) Feature Normalization [ featureNormalize.m ]

 5) Computing Cost (for multiple variables) [ computeCostMulti.m ]

 6) Gradient Descent (for multiple variables) [ gradientDescentMulti.m ]

 7) Normal Equations [ normalEqn.m ]

VII. Regularization (Week 3)  Programming Exercise 2

1) Sigmoid Function  [ sigmoid.m ]

Octave expression:

 2) Logistic Regression Cost [ costFunction.m ]

Vectorization Implementation：

 3) Logistic Regression Gradient [ costFunction.m ]

 4) Predict [ predict.m ]

5)  Regularized Logistic Regression Cost [ costFunctionReg.m ]

not including theta(0) , same as theta(1) in Octave

 6) Regularized Logistic Regression Gradient [ costFunctionReg.m ]

reset the value of grad(1) in line two

5 and 6 can also represented as follows:

VIII. Neural Networks: Representation (Week 4)  Programming Exercise 3

 1) Vectorized Logistic Regression  [ lrCostFunction.m ]

2) One-vs-all classifier training [ oneVsAll.m ]

 3) One-vs-all classifier prediction [ predictOneVsAll.m ]

4) Neural network prediction function [ predict.m ]

IX. Neural Networks: Learning (Week 5)  Programming Exercise 4

1) Feedforward and Cost Function [ nnCostFunction.m ]

 hTheta(x) is the vector with K-dimension

 

 2) Regularized Cost Function [ nnCostFunction.m ]

3) Sigmoid Gradient [ sigmoidGradient.m ]

 4) Neural Network Gradient (Backpropagation) [ nnCostFunction.m ]

cost lots of time.....

 5) Regularized Gradient [ nnCostFunction.m ]

 X. Advice for Applying Machine Learning (Week 6)  Programming Exercise 5

 1) Regularized Linear Regression Cost Function [ linearRegCostFunction.m ]

 2) Regularized Linear Regression Gradient [ linearRegCostFunction.m ]

 3) Learning Curve [ learningCurve.m ]

 4) Polynomial Feature Mapping [ polyFeatures.m ]

 5) Validation Curve [ validationCurve.m ]

not including regularization

 XII. Support Vector Machines (Week 7)  Programming Exercise 6

 1) Gaussian Kernel [ gaussianKernel.m ]

 2) Parameters (C, sigma) for Dataset 3 [ dataset3Params.m ]

calcate the parameters based on training set for every C and sigma pair. Then calcate the cross validation error based on cross validation set for every pair. Choose the minimum pair for cross validation error.

 3) Email Preprocessing [ processEmail.m ]

 4) Email Feature Extraction [ emailFeatures.m ]

 XIV. Dimensionality Reduction (Week 8)  Programming Exercise 7

 1) Find Closest Centroids (k-Means) [ findClosestCentroids.m ]

 2) Compute Centroid Means (k-Means) [ computeCentroids.m ]

vectorization implementation

 3) PCA [ pca.m ]

4) Project Data (PCA) [ projectData.m ]

vectorization implementation

 5) Recover Data (PCA) [ recoverData.m ]

XVI. Recommender Systems (Week 9)  Programming Exercise 8

 1) Estimate Gaussian Parameters [ estimateGaussian.m ]

 2) Select Threshold [ selectThreshold.m ]

3) Collaborative Filtering Cost [ cofiCostFunc.m ]

4) Collaborative Filtering Gradient [ cofiCostFunc.m ]

5) Regularized Cost [ cofiCostFunc.m ]

 6) Regularized Gradient [ cofiCostFunc.m ]