Akansha Yadav breaks down the confusion matrix, a table that shows how a machine learning model's predictions compare to actual results. The matrix has four cells: true positives, true negatives, false positives, and false negatives. It reveals not just accuracy but what kinds of mistakes the model makes.
Confusion metrics. Confusion matrix is the complete picture of model performance. If you wanted to evaluate your model, then confusion matrix is the one of the metrics using which you can analyze your model performance. So basically it will give you one graph where there will be 4 columns. One is TP, FN, TN, FP. So basically on the row side— on the column side it is predicted class, and on the row side it is actual class. Positive-negative? Suppose the actual class is positive and the predicted class is also positive, then we call it true-positive. And when actual class is positive and the predicted class is negative, then it's false-negative. When actual class is negative and the predicted class is positive, then it is false-positive. And when actual class is negative and predicted class is negative, then it is true-negative. So this gives you the entire performance of the model— that how many values are there which are true-positive, means which are evaluated positively and are
