This slecture will be reviewed by Yu Liu.
And here the review goes:
This slecture discussed general procedures of Bayes Classifier in two-class scenario under Gaussian assumption. First it derived the log-discriminant function according to Bayes rule. Next it introduced density estimation technique in general and showed an example of using maximum likelihood estimation (MLE) to estimation the mean and variance of Gaussian data. Finally an experiment was performed to show Bayes classifier in practice. In the experiment MLE was applied to the Gaussian training data for parameter estimation. After that, the estimated parameters were used to classify the testing data with Bayes rule.
What I found good in this slecture is that the idea and the whole structure were quite clear and the experiment was fairly illustrative. However, there are a few things that could be improved: i) In section 2 the first equation is not correct. The right-hand side should be divided by Prob(x) according to the property of conditional probability. ii) The vertical axis of Fig. 2 is just labeled “histogram.” I would suggest using “Number of trials” instead for better understanding. iii) The word “trail” in the text should be “trial.”
