liveBook · Manning

Objective function

As explained in Build AI Drug Discovery Pipelines

A cost function measures aggregate prediction error across multiple data instances, with lower values representing better performance.

An objective function, or cost function, aggregates prediction error across training examples. A loss measures error for an individual example; the cost combines those losses and is minimized by adjusting model weights. section 3.3.1 section 4.2.1

For linear regression predicting solubility, mean squared error averages the squared differences between predicted and known values. Errors farther from the fitted line receive greater penalties, and the model searches for weights that minimize this cost. section 4.2.1

MSE(X, hw)=1n∑j=1n(hw(xj)-yj)2

The equation expresses mean squared error as the average of squared prediction errors across the dataset. section 4.2.1

How Build AI Drug Discovery Pipelines covers it

How deep it goes
Defines aggregate error and works through mean squared error for linear regression.
Examples
Uses mean squared error to fit a solubility regression line and penalize distant residuals.
What you can do after
Apply mean squared error to choose linear-regression weights that fit training values.
How it connects
Connects loss functions, learning routines, model weights, and linear regression.

Also explained in

Compare how 42 books explain objective function →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Build AI Drug Discovery Pipelines. The section links point to the parts of the book it draws on.