Okay. So, let's say I want to do some of the model evaluation. So, for all these classification or these classification models, I can use this confusion matrix to let's say evaluate the accuracy or precision and recall model. So, for classification model, it can be all those our neural network KNN Naive Bayes decision tree. So, for classification model, I can create a confusion matrix using something like this confusion matrix. Then I put in a prediction then I put in a test set then a test our species Okay, so, bodies neural networks So, I create a model to let's say, predict a species based on other other variables.

And then I use a training set and the method is neural network here, I used a model to predict a destiny inside here and the type is probability and I have all these prediction here. So predictions here, show me all the probability I need to remove these. Because for confusion matrix, I cannot put in the audience our probability. So in confusion matrix, I will provide a prediction and then a test set species value. So these are test sets. Our species value is actually the actual Why are the extra variable and then this is the prediction of the variable for this species.

So that's a species is actually the actual value for these species variable prediction is the predicted values for these species variable. So, we this confusion matrix I will put in a prediction and then that test set our species here. So I control a and then run. So this is the confusion matrix. So we have the prediction and we have the actual value or the reference here. Then this is the accuracy then 95% ci now No information rate the p value and then we have the combined p value are any here we have other statistic by cross here, sensitivity specificity law statistic here.

So, this is how we can create these complete confusion matrices ah to get the accuracy of our model. So, we can use this confusion matrix for those our Naive Bayes neural network decision tree KNN od is our classification algorithm. So, for prediction that will be the simple linear regression or this So, prediction is to predict less a numeric or continuous variable. Then all these are classification we predict These are categorical variable okay. So we go into the evaluation of these our prediction model in the next lecture.