What is a confusion matrix in class 10 AI?
It is a two by two table that compares a classification model's predictions with reality. Its four cells count true positives, false positives, false negatives and true negatives. Accuracy, precision, recall and F1 score are all calculated from those four counts, which is why Unit 3 numericals start by filling the matrix.
What is the accuracy formula in class 10 AI?
Accuracy = (TP + TN) / (TP + TN + FP + FN), the correct predictions divided by all predictions. Multiply by 100 for a percentage. For example, 86 correct out of 100 gives 0.86, or 86%. Error rate is the remaining share: (FP + FN) divided by the total.
How do you calculate precision and recall?
Precision = TP / (TP + FP): of everything the model labelled positive, the share that really was. Recall = TP / (TP + FN): of everything that really was positive, the share the model found. Both use TP on top; only the second number below the line differs.
How do you find the F1 score in class 10 AI?
F1 = 2 x precision x recall / (precision + recall). Find precision and recall first, keep three decimals, then substitute. A quick check is 2TP / (2TP + FP + FN), which gives the same value. Never take the plain average of precision and recall: that is a different, wrong number.
How do I make a confusion matrix from a paragraph?
Name the positive class first. Then find TP (predicted yes, really yes), use the predicted-yes total to get FP, the real-yes total to get FN, and subtract all three from the grand total for TN. Add the four cells to confirm they match the number of cases before calculating.
Is accuracy enough to evaluate a model?
No. When one class is rare, a model that always predicts the common class can score 99% accuracy and still find nothing. Two models with the same accuracy can also make very different mistakes. Precision, recall and F1 show what accuracy hides, and Unit 3 numericals often test this.
What is the difference between a false positive and a false negative?
A false positive is a yes prediction that turns out to be no, such as a healthy person flagged as ill. A false negative is a no prediction that turns out to be yes, such as an ill person told they are fine. Which one matters more decides whether precision or recall is the key metric.
How many decimal places should I give?
Two decimal places is the usual safe choice unless the question asks otherwise, and a percentage is acceptable where it fits. Keep more decimals in your working and round only on the final line, because rounding precision and recall early can shift the F1 score.
How many marks do confusion matrix questions carry?
Evaluating Models carries 10 theory marks in CBSE 417, and matrix numericals appear both as one-mark objective items and inside longer case-based questions. The exact split for the 2027 paper will be clear only when CBSE releases the 2026-27 sample paper, so practise both shapes.
Do your classes include confusion matrix practice?
Yes. Matrix numericals are taught and practised in the Monday and Wednesday 9:00 PM IST batch and in personalised one to one, with working marked step by step. The monthly fee shown on this page, in your own currency, covers it. There is no admission fee and the first class is a free demo.
