Overview / Usage

Online learning is a method of machine learning where data becomes available in sequential order and is used to update our best predictor for future data at each step as opposed to batch learning techniques which generate the best predictor by learning on the entire training datasets at once.

1. Online learning is used when it is computationally infeasible to train the entire datasets, requiring the need for algorithms that can process data that are too large to fit in the computer memory at a time.

2. It is used when it is necessary for the algorithm to dynamically adapt to new patterns in the data, or when the data itself is generated as a function in time, e.g stock price prediction

The general usage of this model is to implement incremental machine learning with hybrid feature engineering techniques.

Methodology / Approach

1. We scrapped a news website for different category of news using beautiful soup package in python.

2. We perform text preprocessing such as :

• stop words removal
• type casting
• stemming,
• e.tc

with feature extraction technique using tf-idf bag of words.

1. The output feature representation was passed to different feature engineering techniques such as :

• Chi-Square
• Mutual Information Gain
• PCA

for dimensionality reduction.

​

Technologies Used

Python3

Pandas

Numpy

Matplotlib

Sklearn

Repository

https://github.com/princesegzy01/incremental-machine-learning-techniques-