AI Basics

The training data is assumed to be constant while weights of a Network are said to be variable.

• Machine learning
• Supervised algorithms:
• Classification
• Regression
• Algorithms:
  • Linear Regression
  • Logistic Regression
  • SVM
  • Decision Tree
  • K-Nearest Neighbour
  • Random Forest
  • XG-Boost
• UnSupervised algorithms:
  • CLustering
    • Agglomerative/Hierarchical
    • Point Assignment:
      • K-Means clustering
    • Density Based
  • PCA
  • Autoencoders
• Evaluation metrics:
  • Precision, Recall, Accuracy, ROC AUC Curve, Precision-Recall curve, Confusion matrix.
• Training methodologies:
  • Cross Validation
  • Avoiding Overfitting

In Machine learning, your data consists of Dependent and Independent variables which is also known by other names:

Data is of 2 types at High-level

If response variable is “qualitative - classification” , “quantitative - regression”.

Irrespective of whether the Quantitative response variable is continuous or discrete, regression method is more applicable. Such as, “number of cars”.

However, if response variable can be categorized into certain range or categories like “low”,”medium”, or “high”, then Classification method is more suitable.

If predictor variable is qualitative , then they need to be properly coded before the analysis is performed.

https://www.youtube.com/watch?v=tFWsuO9f74o