WQU-DS-Unit-2

This repo contains all the files material releated to WorldQuant University's Data Science Summer 2020 Session Unit 2: Machine Learning and Statistical Analysis

Syllabus

1. Introduction to machine learning and Scikit-learn API
2. Regression, classification, & model selection (miniproject: ml)
3. Feature engineering
4. NLP and dimension reduction (miniproject: nlp)
5. KNeighbors, clustering and ensemble models
6. Support vector machines
7. Time series analysis and anomaly detection
8. Clustering

Chapter Content Videos

Ch1. Introduction to machine learning and Scikit-learn API

• Introduction to Machine Learning
• 1.2.1 Intro to Scikit-learn
• 1.2.2 Predictors
• 1.2.3 Transformers and Pipeline
• 1.2.4 Feature Unions
• 1.2.5 Custom Transformers
• 1.2.6 Custom Predictors
• 1.2.7 Exercise Distance Transformer
• 1.2.8 Exercise Majority Classifier
• 1.3.1 Persisting Your Model
• 1.3.2 Common Mistakes

Ch2. Regression, classification, & model selection

• 2.1.1 Regression Metrics
• 2.2.1 Linear Regression Intro
• 2.2.2 Gradient Descent and Huber Loss
• 2.2.3 Multivariate Regression
• 2.2.4 Feature Importance
• 2.3.1 Classification Metrics
• 2.3.2 Probabilistic Models and Metrics
• 2.3.3 Logistic Regression
• 2.3.4 Multiclassification
• 2.4.1 Model Selection
• 2.4.2 Intro to Decision Trees
• 2.4.3 Underfitting and Overfitting
• 2.4.4 GridSearchCV
• 2.4.5 Comparing Two Models
• 2.5.1 Imputation
• 2.5.2 Categorical Data
• 2.6.1 GridsearchCV and Pipelines 1
• 2.6.2 GridsearchCV and Pipelines 2
• 2.6.3 RandomizedSearchCV

Ch3 Feature Enginerring & KNN

• 3.1.1 Feature Engineering and Extraction
• 3.1.2 Feature Transformation
• 3.1.3 Curse of Dimensionality
• 3.1.4 Regularization
• 3.1.5 Multicollinearity and PCA
• 3.1.6 Ensemble Models
• 3.2.1 Bias and Variance
• 3.2.2 Learning Curves
• 3.3.1 Intro KNN
• 3.3.2 KNN Bias and Variance
• 3.3.3 KNN Time Complexity
• 3.3.4 KD Trees and Weights
• 3.4.1 Intro to NLP
• 3.4.2 Spacy
• 3.4.3 Obtaining a Corpus
• 3.4.4 Bag of Words Model
• 3.4.5 Hashing Vectorizer
• 3.4.6 TF-IDF
• 3.4.7 Improving Signal
• 3.4.8 N-grams and Similarity
• 3.4.9 Word Usage Classifier
• 3.4.10 Exercise I
• 3.4.11 Exercise II
• 3.4.12 Exercise III and IV

Ch4 Decision Trees & Gradient Boosting

• 4.1.1 Intro to Decision Trees
• 4.1.2 Tree Error Metrics
• 4.1.3 Trees for Regression
• 4.1.4 Training Trees and Hyperparameters
• 4.1.5 Geometric Interpretation and Time Complexity
• 4.1.6 Time Complexity Continued
• 4.1.7 Random Forests
• 4.1.8 Extreme Random Forests
• 4.1.9 Gradient Boosting Trees I
• 4.1.10 Gradient Boosting Trees II
• 4.1.11 Feature Importance
• 4.1.12 Exercises

Ch5 SVM & Clustering

• 5.1.1 Intro to SVM
• 5.1.2 Largest Margin Classifier
• 5.1.3 Soft Margin Classifier
• 5.1.4 SVM Kernels
• 5.1.5 SVM vs Logistic Regression
• 5.1.6 SVM Regression
• 5.1.7 SVM Lagrangian Dual
• 5.1.8 Kernel Trick
• 5.1.9 SVM Time Complexity and Multiclass
• 5.1.10 SVM Tuning Kernels Exercise Part I
• 5.1.11 SVM Tuning Kernels Exercise Part II
• 5.1.12 SVM Kernel Approximations
• 5.1.13 SVM Online Learning
• 5.1.14 SVM Online Learning Pipeline
• 5.2.1 Intro to Clustering
• 5.2.2 Metrics for Clustering
• 5.2.3 KMeans Clustering
• 5.2.4 Elbow Plots
• 5.2.5 Gaussian Mixture Models
• 5.2.6 Choosing Cluster Based on Silhouette
• 5.2.7 GMM Choosing Number of Components

Ch6 Time Series Analysis & Dimensionality Reduction

• 6.1.1 Intro to Time Series
• 6.1.2 Crossvalidation in Time Series
• 6.1.3 Stationary Signal
• 6.1.4 Modeling Drift
• 6.1.5 Fourier Transforms Part I
• 6.1.6 Fourier Transforms Part II
• 6.1.7 Fourier Components in our Model
• 6.1.8 Modeling Noise
• 6.1.9 Moving Statistics
• 6.1.10 Full Model
• 6.1.11 ARMA and ARIMA
• 6.1.12 AR Example
• 6.2.1 Intro to Dimension Reduction
• 6.2.2 Math of Projections
• 6.2.3 PCA
• 6.2.4 PCA in Scikit Learn
• 6.2.5 PCA Implementation Details
• 6.2.6 Choosing the Number of Components
• 6.2.7 Truncated SVD
• 6.2.8 NMF
• 6.2.9 Using PCA with Supervised ML
• 6.2.10 PCA for Visualization
• 6.2.11 NMF Exercise Part I
• 6.2.12 NMF Exercise Part II
• 6.2.13 Variants of PCA

Ch7 Anomaly Detection

• 7.1.1 Intro to Anomaly Detection
• 7.1.2 One class SVM
• 7.1.3 Isolation Forest
• 7.1.4 Comparison Between One-class SVM and Isolation Forest
• 7.1.5 Intro to Case Study
• 7.1.6 Initial Baseline Model Part I
• 7.1.7 Initial Baseline model Part II
• 7.1.8 Full Baseline Model
• 7.1.9 Z-score Detection
• 7.1.10 Rolling Z-score Detection
• 7.1.11 Using External Features Initial Model
• 7.1.12 Using External Features Tuning the Model
• 7.1.13 Packaging the Time Series Anomaly Detector

Ch8 Model Deployment

• 8.1.1 Model Considerations
• 8.1.2 Model Development
• 8.1.3 Flask App Local Development
• 8.1.4 GET requests
• 8.1.5 Making GET Requests with Model
• 8.1.6 Using our Model with Twitter Web API
• 8.1.7 POST Requests and Flask Templates
• 8.1.8 Preparing for Deployment to the Web
• 8.1.9 Deploying our App to the Web with Heroku
• 8.2.1 Rethinking Model Tuning
• 8.2.2 Intro to Bayes Theorem
• 8.2.3 Bayesian Inference
• 8.2.4 Bayesian Optimization
• 8.3.1 End of Course Material

Office Hours

• LIVE Streaming Office Hours
• LIVE Streaming Office Hours
• LIVE Streaming Office Hours
• LIVE Streaming Office Hours
• OFFICE HOURS PLAYLIST