Machine Learning Week 1 Assignment

Questions and Answers

What is the primary purpose of using a validation dataset in machine learning?

To evaluate the performance of the model after training.

To tune the hyperparameters of the machine learning model. (correct)

To collect data for training the model.

To select features for the model.

Which statement correctly describes models that underfit the training data?

They exhibit low bias and high variance.

They demonstrate high bias and low variance. (correct)

They usually have a good generalization to new data.

They capture the complex patterns in the data.

Which of the following options represents a continuous feature?

Height of a person. (correct)

Mother tongue of a person.

Preferred mode of transportation.

Number of languages spoken.

In the context of bias and variance, which of the following statements is true?

Overfitting leads to low bias and high variance. Signup and view all the answers

Which of the following is an example of a categorical feature?

Mother tongue of a person. Signup and view all the answers

What is the precision if True Positives (TP) are 50 and False Positives (FP) are 100?

33.33% Signup and view all the answers

Which of the following is an unsupervised learning problem?

Grouping customers based on purchasing behavior. Signup and view all the answers

Which machine learning paradigm is best suited for learning complex strategies in a game with no prior knowledge?

Reinforcement learning Signup and view all the answers

How many different Boolean functions can be created with 3 features?

2^3 Signup and view all the answers

What is the role of a validation dataset in machine learning?

To evaluate the model's performance after training. Signup and view all the answers

What is the recall if True Positives (TP) are 50 and False Negatives (FN) are 250?

16.67% Signup and view all the answers

Which of the following options describes supervised learning?

Classifying images using labeled datasets. Signup and view all the answers

What defines a successful reinforcement learning strategy in a game environment?

Maximizing total rewards through learned actions. Signup and view all the answers

Which of the following describes a classification problem?

Determining whether an email is spam or not Signup and view all the answers

Which algorithm is used for clustering tasks?

K-Means Signup and view all the answers

What is the main goal of regression tasks in machine learning?

Predicting a continuous outcome Signup and view all the answers

Which option reflects a non-supervised learning task?

Clustering customers based on purchasing behavior Signup and view all the answers

What type of problem is represented by predicting the stock market price?

Regression Signup and view all the answers

Which of the following best defines precision in the context of a spam detection system?

The ratio of true positives to all predicted positives Signup and view all the answers

Which task is least appropriate for machine learning?

Finding the shortest path in a network Signup and view all the answers

How is recall calculated in a classification context?

True Positives / (True Positives + False Negatives) Signup and view all the answers

Study Notes

Classification Tasks

Classification tasks involve discrete class outputs, like detecting pneumonia from chest X-ray images.
Other examples like predicting price or temperature are regression tasks as they deal with continuous outputs.

Supervised Learning Types

Supervised learning requires target values for training through Classification or Regression.
Clustering is unsupervised learning and does not rely on labeled data.

Machine Learning Suitability

Not all tasks are suitable for machine learning, such as finding the shortest path in graph theory.
Machine learning is more suited for predictive tasks like stock price prediction or spam detection.

Spam Detection Metrics

In spam detection, precision and recall are key metrics.
For a system detecting 150 spam emails with 50 true positives, precision equals 33.33% and recall equals 16.66%.

Supervised Learning Problems

Predicting disease from blood samples and face recognition are examples of supervised learning problems.
Grouping students based on features is considered an unsupervised learning problem.

Reinforcement Learning

In complex scenarios, like unfamiliar games, reinforcement learning is optimal for developing strategies.
It focuses on maximizing rewards based on the game's outcome.

Boolean Functions

The number of possible Boolean functions with N features is 2^(2^N).
Each feature can be either True (1) or False (0), leading to diverse combinations.

Validation Dataset Use

A validation dataset is critical for tuning hyperparameters in machine learning models.
It is not used for training or direct performance evaluation, which are different processes.

Bias and Variance

Overfitting leads to low bias and high variance, while underfitting results in high bias and low variance.
Understanding these concepts helps in developing models that balance performance.

Categorical Features

Categorical features are data types that can be grouped, such as a person's mother tongue.
Continuous variables, like height or price, differ from categorical data types.

Studying That Suits You

Use AI to generate personalized quizzes and flashcards to suit your learning preferences.

Description

This quiz consists of 10 multiple-choice questions based on the introductory concepts of Machine Learning. It is part of the NPTEL Online Certification Courses offered by the Indian Institute of Technology Kharagpur. Test your understanding and earn marks with each correct answer!