Data Science Part III (Model Building Process)

[8bitzz]

Model Building Process:

1. Feature Engineering
2. Model Selection (Algorithms)
3. Model Training
4. Model Validation and Selection
5. Applying model to unseen data

Feature Engineering

Some techniques:

• Correlation Analysis
• Recursive Feature Elimination
• Leave-one-out
• Leave-K-out
• Select-K-best(Classification)

Source: Scikitlearn

Model Selection (Algorithms)

Supervised Learning

Regression (Prediction)

1. Linear Regression
2. Polynomial Regression
3. Ridge/Lasso Regression

Classification

1. Decision Trees
2. Logistics Regression (BE CAREFUL!!!)
3. Naive Bayes
4. K-NN
5. SVM

Unsupervised Learning

Clustering

1. DBSCAN
2. K-Means
3. Mean-Shift
4. Fuzzy C-Means
5. Agglomerative