1 of 10

ECONOMIC STATISTICS-I�(TYPES OF CLASSIFICATION)

Prepared

By

M.Selvi

Assistant Professor of economics

CPA College,

Bodinayakanur

2 of 10

INTRODUCTION TO CLASSIFICATION�

Definition: Classification is a type of supervised learning where the goal is to predict a categorical label (class) for a given input.

Objective: To group data into classes based on input features.

Examples:

    • Email spam detection (Spam or Not Spam)
    • Disease diagnosis (Healthy or Sick)
    • Sentiment analysis (Positive, Neutral, or Negative)

​

3 of 10

MAIN CATEGORIES

​

    • Binary Classification
    • Multiclass Classification
    • Multilevel Classification

​

4 of 10

BINARY CLASSIFICATION�

Definition: Classification task where there are only two possible outcomes. Example: Email spam detection (Spam vs. Not Spam)Algorithms Used: Logistic Regression

Decision Trees

Support Vector Machines (SVM)

K-Nearest Neighbors (KNN)

​

5 of 10

MULTICLASS CLASSIFICATION�

Definition: Classification task with more than two classes (but each instance belongs to exactly one class).

Example: Classifying types of flowers (e.g., Rose, Tulip, Daisy)

Algorithms Used:

    • Decision Trees
    • Random Forest
    • Naive Bayes
    • Neural Networks

​

​

6 of 10

MULTILABEL

Definition: Each instance can belong to multiple classes simultaneously.

Example: A movie could belong to both "Action" and "Comedy" genres.

Algorithms Used:

    • Binary Relevance
    • Classifier Chains
    • Neural Networks (for deep multilabel learning)

7 of 10

ADDITIONAL CLASSIFICATION TYPES�

Imbalanced Classification:

    • When one class is significantly more frequent than the other.
    • Solutions: Resampling, Synthetic Data Generation (e.g., SMOTE), Cost-sensitive learning.

Ordinal Classification:

    • When labels have a natural order (e.g., Low, Medium, High).

Hierarchical Classification:

    • When labels are structured in a hierarchy or tree (e.g., Animal -> Mammals -> Dogs).

​

8 of 10

COMMON CLASSIFICATION ALGORITHMS�

Overview of Common Algorithms:

    • Logistic Regression
    • Naive Bayes
    • Decision Trees & Random Forests
    • K-Nearest Neighbors (KNN)
    • Support Vector Machines (SVM)
    • Neural Networks
    • Gradient Boosting Machines (GBM)

​

9 of 10

CHALLENGES IN CLASSIFICATION�

Over fitting: Model performs well on training data but poorly on unseen data.

Under fitting: Model is too simple to capture the underlying patterns in data.

Class Imbalance: One class has far fewer samples than others.

Noise and Outliers: Irrelevant data points that disrupt the model.

​

10 of 10

THANK YOU