摘要： 1. Content based Problem formulation Content Based Recommendations:2. collaborative filtering algorithm 阅读全文

摘要： 1. Algorithm2. evaluating an anomaly detection system3. anomaly detection vs supervised learning4. choose what features to use. - choose the features xi which hist(xi) is like gaussian shape, or transfer xi such as log(xi+c) to make hist(xi) to be like gaussian shape. - if anomaly case's feature 阅读全文

摘要： 1. Principal component analysis algorithmdata preprocessing2. choosing the number of principal components3. reconstruction from compressed representation4. Application of PCA - compression - reduce memory/dist needed to store data - speed up learning algorithm - visualizationbad use of PCA: to pre.. 阅读全文

摘要： 1. K-means algorithm2. K-means optimization objective3. Random initialization 阅读全文

摘要： 1. SVM hypothsis2. large margin classification3. kernals and similarityiff1 = 1;if x if far from l^(1), f1 = 04. SVM with kernels5. SVM parameters6. Multi-class classification7. Logistic regression vs SVMs 阅读全文

摘要： 1. advantage: when number of features is too large, so previous algorithm is not a good way to learn complex nonlinear hypotheses.2. representation"activation" of unit i in layer jmatrix of weights controlling function mapping from layer j to layer j+13. samplewe have the neural expression 阅读全文

摘要： 1. Underfit = High bias Overfit = High varience2. Addressing overfitting: (1) reduce number of features. Manually select which features to keep. Model selection algorithm disadvantage: throw out some useful information (2) Regularization Keep all the features, but reduce magnitude/valu... 阅读全文

摘要： 1. Decision boundarywhen hθ(x) > 0, g(z) = 1; whenhθ(x) < 0, g(z) = 0.so the hyppthesis is:2. cost functionto fit parametersθ:to make a prediction given new x:Output 3. Gradient DescentRepeat { (simultaneously update allθj)} 阅读全文

摘要： 寻找比n小的所有质数的方法。2是质数， 2*i都是质数，同样3是质数，3*i也都是质数代码如下1 int n;2 vector prime (n+1, true);3 prime[0] = prime[1] = false;4 for (int i=2; i<=n; ++i)5 if (prime[i])6 if (i * 1ll * i <= n)7 for (int j=i*i; j<=n; j+=i)8 prime[j] = false; 阅读全文

摘要： 1. notation:n = number of featuresx(i) = input (features) of ithtraining example= value of feature j inithtraining example2. Hypothesis:3. Cost function:4. Gradient descent:Repeat { }substituting cost function, thenRepeat { (simultaneously updateθj for j = 0, ... n)}5. Mean normalizationreplace ... 阅读全文