Class 10 AI tuition in Pune, for CBSE students, without a trip across the city.

We found no official figure for how many Pune CBSE schools offer 417, so we print none. Not all CBSE schools offer it; the timetable will show whether your child has it.

1. Accuracy

   (TP + TN) / total = (30 + 74) / 120 = 104 / 120 = 0.87 (0.8667). The model is wrong on 16 days.

2. Precision

   TP / (TP + FP) = 30 / 40 = 0.75. When it says rain, it is right three times in four.

3. Recall

   TP / (TP + FN) = 30 / 36 = 0.83 (0.8333). Of the 36 rainy days, it caught 30.

4. F1 score

   2 x P x R / (P + R) = 2 x 0.75 x 0.8333 / 1.5833 = 0.79 (0.7895). The plain average of P and R, 0.7917, happens to round to the same 0.79 here, but on other numbers it does not, so always use the formula.

5. (c) Precision or recall?

   Recall. A missed rainy day (FN) means children and equipment caught in a downpour; a false alarm (FP) only moves the event indoors. A good answer names the costly error and links it to the metric.

One more line we always add in class: a lazy model that says "no rain" every day would be right on the 84 dry days, an accuracy of 0.70, with a recall of zero. That is why accuracy alone is never the whole answer. More numericals of this kind are on the confusion matrix numericals page, and the ideas behind them on the evaluating models page.

1. Pick early, in class

   The idea is fixed in the first weeks so it can grow alongside the units, one project-cycle stage at a time.

2. Write it up as you go

   Each stage, from problem scoping to evaluation, is recorded the week it is done, so there is no write-up panic at the end.

3. Explain it out loud, twice

   Each student presents their project to the class and answers follow-up questions, then does it again a fortnight later. More ideas are on the project ideas page, and the oral questions on the viva questions page.