1. Predicting Housing Prices
1.1. Supervised learning
1.1.1. Regression
1.1.1.1. Models
1.1.1.1.1. Linear regression
1.1.1.1.2. Regularization
1.1.1.2. Algorithms
1.1.1.2.1. Gradient descent
1.1.1.2.2. Regression Model
1.1.1.2.3. Fitting in a regression model:
1.1.1.3. Concepts
1.1.1.3.1. Loss Functions
1.1.1.3.2. Bias-Variance tradeoff
1.1.1.3.3. Cross Validation
1.1.1.3.4. Sparsity
1.1.1.3.5. Overfitting
1.1.1.3.6. Model Selection
1.1.1.3.7. Training, test and validation set
2. Loan Safety Sentiment Analysis
2.1. Supervised Learning
2.1.1. Classification
2.1.1.1. Models
2.1.1.1.1. Linear classifiers (logistic regression)
2.1.1.1.2. Multiclass classifiers
2.1.1.1.3. Decision Trees
2.1.1.1.4. Bagging
2.1.1.1.5. k-nearest neighbors classification
2.1.1.1.6. Boosted decision trees
2.1.1.1.7. random forests
2.1.1.2. Algorithms
2.1.1.2.1. Boosting
2.1.1.2.2. Learning from weighted data
2.1.1.3. Concepts
2.1.1.3.1. Decision boundaries
2.1.1.3.2. Maximum likelihood estimation
2.1.1.3.3. Ensemble methods
2.1.1.3.4. Random forests
2.1.1.3.5. Precision and recall
2.1.1.3.6. Bias and Fairness in ML
3. Image Classification
3.1. Supervised or unsupervised
3.1.1. Deep Learning
3.1.1.1. Models
3.1.1.1.1. Perceptron
3.1.1.1.2. General neural network
3.1.1.1.3. Convolutional neural network
3.1.1.2. Algorithms
3.1.1.2.1. Convolutions
3.1.1.2.2. Backprogation (high level only)
3.1.1.3. Concepts
3.1.1.3.1. Activation Function
3.1.1.3.2. Hidden Layers
3.1.1.3.3. Architecture choices
4. Document Clustering & Analysis
4.1. Unsupervised
4.1.1. Clustering & Retrieval
4.1.1.1. Models
4.1.1.1.1. Clustering
4.1.1.1.2. Mixture Models
4.1.1.1.3. Hierarical clustering
4.1.1.2. Algorithms
4.1.1.2.1. k-means
4.1.1.2.2. k-means++
4.1.1.2.3. Agglomerative
4.1.1.2.4. Divisive Clustering
4.1.1.2.5. Principal Component Analysis (PCA)
4.1.1.3. Concepts
4.1.1.3.1. Clustering
4.1.1.3.2. Dimensionaity Reduction
4.2. Unsupervised learning:
4.2.1. o Can create groups of things that are similar to each other
4.2.2. o Many applications
4.2.3. o Learning things with underlying patterns rather than set outputs that we know
5. Recommender Systems
5.1. Supervised or unsupervised
5.1.1. Recommender Systems & Matrix Factorization
5.1.1.1. Models
5.1.1.1.1. Collaborative filtering
5.1.1.1.2. Popularity
5.1.1.1.3. Nearest User(user-user)
5.1.1.1.4. Peopl who bought this also bought Item-item
5.1.1.1.5. classification
5.1.1.1.6. Matrix factorization
5.1.1.2. Algorithms
5.1.1.2.1. Coordinate descent
5.1.1.3. Concepts
5.1.1.3.1. Matrix completion
5.1.1.3.2. cold-start problem
5.1.1.3.3. co-occurance matrix
5.1.1.3.4. Jaccard similarity