python ai

b — the number of nearest neighbors

What does the “k” in k-Nearest Neighbor represent?

a — stored training samples

What does a nearest neighbor classifier mainly use to classify a new point?

b — distance

What must be calculated between a test point and each training point?

b — it may include points from other classes

What happens when k is very large?

b — if it looks like something, it is likely that thing

Which simple idea describes the basic concept of nearest neighbor classification?

c — a method that repeatedly asks simple yes/no questions to make predictions

What is a Decision Tree?

c — supervised learning

According to the slides, what type of learning does a Decision Tree use?

c — the data is split into smaller and more uniform groups

In a Decision Tree, what happens each time a question is asked?

c — Boolean (yes/no) questions

In the commute example, which type of questions are used in the tree?

a — to choose attributes that give the cleanest splits

What is the purpose of selecting attributes when building a Decision Tree?

a — the posterior probability of a hypothesis

What does Bayes Theorem help us compute in classification?

d — attribute values are conditionally independent given the class

What is a key assumption of the Naive Bayes Classifier?

d — probability of a hypothesis before seeing any data

What does the prior probability represent?

d — a probability distribution over all possible classes

What does the Naive Bayes Classifier output?

b — it is simple and works well even with small datasets

What is one advantage of the Naive Bayes Classifier mentioned in the slides?

a — to evaluate how well a model predicts future data

What is the main purpose of cross validation?

a — it is randomly split into training and test sets

In the test-set (hold-out) method, what typically happens to the dataset?

a — it wastes data because only part of the dataset is used for training

What is one downside of the test-set method mentioned in the slides?

c — all points except one

In Leave-One-Out Cross Validation (LOOCV), how many points are used for training each time?

b — average the errors across all folds

In k-fold Cross Validation, what do we do after computing the error for each fold?

a — to maximize the evaluation function by moving to better states

What is the main goal of hill-climbing?

b — it can get stuck in a local optimum

What is a common problem with basic hill-climbing?

c — it picks a random move from the moveset

How does randomized hill-climbing differ from normal hill-climbing?

b — accept worse moves with some probability

What is the key idea of simulated annealing?

b — natural selection and evolution

What biological process inspires the genetic algorithm?

a — to load the Iris dataset

What is the purpose of the following code: from sklearn.datasets import loadiris; iris = loadiris()?

a — train_test_split

Which function is used to split the data into train and test sets?

a — test_size

Which parameter in traintestsplit specifies the proportion of the test set?

a — create an object with LogisticRegression() and then call fit()

Which of the following is the correct way to create a classification model in scikit-learn?

b — to train the model on the training data

What is the role of model.fit(Xtrain, ytrain)?

c — to use the train/test splitting function traintest_split

What is the purpose of the following import statement: from sklearn.modelselection import traintestsplit?