Interview Machine Learning Questions Set 2(1 to 10)

##datascience ##machinelearning ##statistics ##mlalgorithms

Akash Borgalli Dec 30 2021 · 1 min read
Share this

1. How you can define Machine Learning?

Ans: With the help of some equations in an algorithm when we are able to find out a relation or pattern between data that is called ML. It just finds a value of constants(parameters).

2. What do you understand about Labelled training datasets?

Ans: The training dataset has an output feature meaning training data holds the feature that it needs to predict.

3. What are the 2 most common supervised ML tasks you have performed so far?

Ans: I have worked on Linear Regression and Logistic Regression tasks.

4. What kind of Machine learning algorithm would you use to walk robots in various unknown areas?

Ans: I will use ComputerVision and Reinforcement Learning to make robots walk in the various unknown regions.

5. What kind of ML algorithm you can use to segment your user into multiple groups?

Ans: I can use clustering ML Algorithms that come under unsupervised Machine Learning which are K-Means, Hierarchical clustering, DBSCAN in such cases.

6. What type of learning algorithm is realized on similarity measure to make a prediction?

Ans: Its instance-based algorithm that realizes based on similarity measure to make a prediction.

Q. What kind of learning algorithm makes predictions using a similarity measure?

Ans: Instance Based Algorithms :  They memorize and then apply(Instead of performing explicit generalization, it compares new problem with instances in training which are stored in memory which is also called as lazy learning or memory based learning.For Example:- Spam mails. There are 3 types of instance based learning

  • Lazy Learners(KNN)
  • Radial based functions(RBF(related to weights))
  • Case-based reasoning.
  • 7. What is an online learning system?

    Ans: It’s a system in which the machine tries to learn from small chunks of data which is fed continuously.

    8. What is an out-of-core learning system?

    Ans: It’s a type of system that can handle data that cannot fit into your computer memory. It uses an online learning system to feed data in small bits.

    9. Can you name a couple of ml challenges that you have faced?

    Ans: I have faced challenges when I see a model is overfitted model, under-fitted model, or when it has less data that is not labeled correctly, selecting hyperparameters to tune the model.

    10. Can you please give 1 example of hyperparameter tuning wrt some classification algorithm?

    Ans. I can use hyperparameter tuning options like GridSearchCV or Random Search CV or Optuna and the parameters that I will use if I am using logistic regression would be are the penalty(L1 or L2 regularization), solver, max_iter, random_state.

    Read next