Little Bits of Artificial Intelligence

1. When is MSE Most Relevant?

MSE is best suited for regression problems, where our network’s output is a real-valued number (not a class/category). Here are some example use cases:

• House Price Prediction – Predicting the price of a house based on size, location, etc.
• Stock Price Forecasting – Estimating future stock values.
• Temperature Forecasting – Predicting weather data.
• Sales Forecasting – Predicting future sales based on trends.
• Medical Diagnosis – Predicting blood glucose or blood pressure levels.

Basic Math Behind MSE

MSE stands for Mean Squared Error, and it’s calculated as:

Where:

• yiy_iyi​: actual/true value
• y^i\hat{y}_iy^​i​: predicted value
• nnn: number of samples

   It’s the average of the squared differences between actual and predicted values.

Why square the difference?

• To ensure error is always positive (no canceling out).
• To penalize larger errors more heavily (important in many regression tasks).