AIExplainer

What is bias (math) or bias term?

A constant added to a linear combination of inputs in a machine learning model

In machine learning, the bias term is a constant that is added to the weighted sum of inputs to a model, allowing the model to shift the activation curve to the left or right, which can improve the model's ability to fit the data

Think of the bias term like the intercept in a linear equation, where it shifts the line up or down to better fit the data points

For example, in a model that predicts house prices based on features like number of bedrooms and square footage, the bias term might represent the average price of a house in the area, which is then adjusted based on the input features

The bias term is used in linear regression, logistic regression, and neural networks to improve the model's performance and ability to generalize to new data

One common misconception is that the bias term is only used to 'fix' a model that is not performing well, when in fact it is a fundamental component of many machine learning models

The concept of bias in machine learning has its roots in statistical modeling, where it was used to account for the intercept or constant term in linear regression

intercept constant term offset

Three products for different needs — explore what’s relevant to you.