Describe Core Machine Learning Concepts AI-900 Practice Questions (Page 6)

Questions 26–30

1. 26application · medium

   A team is building a model to predict loan default risk. They have a dataset of 50,000 historical loan records with the outcome already labeled. The team wants to train the model and then tune hyperparameters using a separate subset. Which approach best supports this workflow?

   • ASplit the data into a training set and a validation set, train on the training set, and tune using the validation set
   • BUse all 50,000 records for training and then test on the same data
   • CRandomly select 1,000 records for training and use the rest for validation
   • DTrain on the validation set and validate on the training set
   Reveal Solution Discussion Select an answer first
2. 27foundation · easy

   Which statement correctly describes the typical process of splitting a dataset?

   • AThe dataset is split into a training set and a validation set, and the validation set is used to train the model further.
   • BThe dataset is divided so that the validation set contains only the most difficult examples.
   • CThe dataset is split into two equal halves, and both are used to train the model.
   • DThe dataset is randomly divided into a training set and a validation set, with the training set used to fit the model and the validation set used to evaluate it.
   Reveal Solution Discussion Select an answer first
4. 28foundation · easy

   What does the process of splitting a dataset into training and validation datasets accomplish?

   • AIt allows the model to be evaluated on data it has not seen during training.
   • BIt removes outliers and noisy data from the dataset.
   • CIt increases the size of the dataset by duplicating samples.
   • DIt ensures the model is trained on all available data to maximize accuracy.
   Reveal Solution Discussion Select an answer first
5. 29application · medium

   A healthcare organization wants to predict whether a patient has diabetes based on glucose level, blood pressure, and BMI. They have historical patient records with these measurements and a column indicating diabetes diagnosis (Yes/No). What should be the label in this supervised learning scenario?

   • AThe patient ID column
   • BThe glucose level column
   • CThe BMI column
   • DThe diabetes diagnosis column (Yes/No)
   Reveal Solution Discussion Select an answer first
6. 30application · medium

   A data scientist is building a model to predict customer churn. The dataset has 50,000 rows and is heavily imbalanced: only 5% of customers churned. The scientist wants to ensure the validation set reflects the same proportion of churners as the overall dataset. Which splitting technique should be used?

   • ASplitting by customer ID parity
   • BRandom splitting without stratification
   • CStratified splitting
   • DChronological splitting
   Reveal Solution Discussion Select an answer first