-
1
Show your ID card.
-
2
Explain your notebook/project.
-
3
Explain your notebook as a story.
-
4
Explain your EDA.
-
5
Explain the graphs you created.
-
6
What insights did you draw from the graphs?
-
7
Difference between a bar plot and a histogram.
-
8
What does a pair plot show?
-
9
What is the shaded region in a regression plot?
-
10
Why is the correlation between two features zero?
-
11
How does correlation affect model predictions?
-
12
What does errors='coerce' do while converting dates?
-
13
Explain your preprocessing steps.
-
14
Explain your train-test split.
-
15
What happens if the train-test split ratio changes?
-
16
Explain your encoding.
-
17
Explain your scaling.
-
18
Explain your feature engineering.
-
19
Why did you use ColumnTransformer?
-
20
Why didn't you use Pipeline?
-
21
Have you used user-defined functions?
-
22
Explain your hyperparameter tuning.
-
23
Why did you use RandomizedSearchCV?
-
24
Difference between GridSearchCV and RandomizedSearchCV.
-
25
How many times will RandomizedSearchCV run?
-
26
What scoring metrics can be used in GridSearchCV?
-
27
How do different scoring metrics affect model selection?
-
28
What models did you use?
-
29
Why did you choose those models?
-
30
What is your baseline model?
-
31
Why did you use only boosting models?
-
32
Difference between Random Forest and XGBoost.
-
33
What is Random Forest?
-
34
Why didn't you choose Random Forest?
-
35
What is XGBoost?
-
36
What is LightGBM?
-
37
What are the important LightGBM parameters?
-
38
What is early_stopping?
-
39
After early stopping, when is the best score reported?
-
40
What are min_samples_split and min_samples_leaf?
-
41
If a node has 7 samples, will it split?
-
42
Which of your models are parametric?
-
43
Which of your models are non-parametric?
-
44
Why is Logistic Regression called "Regression" even though it performs classification?
-
45
What is MLPClassifier?
-
46
What is Ridge Regression?
-
47
What is Lasso Regression?
-
48
Which regression technique can be used for feature selection?
-
49
What is Stacking?
-
50
How does stacking work?
-
51
How does a boosting model learn?
-
52
How does a Ridge meta-model combine XGBoost, CatBoost, and LightGBM?
-
53
How did you choose ensemble weights?
-
54
Why did your ensemble use those weights?
-
55
What is weak learner predictive power?
-
56
How much data is used in Bagging?
-
57
What is F1-Score?
-
58
Write the formula for F1-Score.
-
59
Does F1-Score give equal importance to Precision and Recall?
-
60
Why did you use F1 Macro?
-
61
Why didn't you use Accuracy?
-
62
Explain the result comparison graph.
-
63
Explain your model performance comparison.
-
64
Coding: Load the Iris dataset and print the feature matrix.
-
65
Coding: Load the Iris dataset and print the target values.
-
66
Coding: Load the Iris dataset and separate features and target.
-
67
Coding: Load the Diabetes dataset and print the feature matrix.
-
68
Coding: Load the Breast Cancer dataset and print the target names.
-
69
Coding: Load the California Housing dataset and train a HistGradientBoostingRegressor.
-
70
Coding: Draw a histogram using the Diabetes dataset.
-
71
Coding: Plot a bar graph comparing models and their scores.
-
72
Coding: Plot the target variable after applying log transformation.
-
73
Coding: Implement GridSearchCV or RandomizedSearchCV.
-
74
Coding: Import a dummy/sample dataset.
-
75
Coding: Implement Lasso Regression.
-
76
Coding: Print the first few rows of the Iris dataset.
-
77
Coding: Use fit_transform() output as a Pandas DataFrame using set_output(transform="pandas").