Many machine learning and statistical algorithms can be formulated as least square procedures. They are effectively minimising the sum of the squares of the differnce between the result they deliver and the desired result -- called the mean squared error. In some case this is explicit in the way the algorithm is forumulated, for example the backpropagation algorithm is a form of gradient descent that calculates the changes needed to each weight to make the best reduction in the mean square. Simularly, linear rgeression minimises the sum of squares of the residuals (that is the difference between the fitted {[regression line}} and the actual data points).
Used on pages 130, 141