Machine Learning Interview Questions 2025 - Series2

11. 🧠 What is Dimensionality Reduction in Machine Learning?

Dimensionality reduction refers to the process of reducing the number of input variables or features in a dataset. High-dimensional data can be complex, slow to process, and prone to overfitting. Techniques like Principal Component Analysis (PCA) and t-SNE help simplify datasets while preserving their essential patterns.

Benefits:

  • Improves model performance
  • Reduces training time
  • Helps visualize high-dimensional data

12. ⚙️ What Are Hyperparameters? How Are They Different from Parameters?

Hyperparameters are settings that are set before training a model, like learning rate, number of trees in Random Forest, or batch size in neural networks. Parameters, on the other hand, are values learned by the model itself, such as weights in linear regression.

Hyperparameter tuning is crucial for optimizing model performance and can be done using Grid Search or Random Search.

13. 🧪 What is Model Validation?

Model validation checks how well your machine learning model performs on unseen data. It’s essential to ensure your model isn’t just memorizing the training data.

Common Validation Techniques:

  • Hold-out Validation
  • k-Fold Cross Validation
  • Leave-One-Out Cross Validation (LOOCV)

14. 📉 What is Gradient Descent?

Gradient Descent is an optimization algorithm used to minimize the cost or error in a model by adjusting parameters like weights. It works by computing the gradient (slope) of the cost function and moving in the direction of steepest descent.

Types: Batch, Stochastic, and Mini-batch Gradient Descent.

15. 🔁 What is the Difference Between Bagging and Boosting?

Both are ensemble learning techniques, but they work differently:

  • Bagging (Bootstrap Aggregating): Builds multiple models independently and combines their output (e.g., Random Forest).
  • Boosting: Builds models sequentially, where each model corrects the errors of the previous one (e.g., XGBoost, AdaBoost).

16. 📊 What is ROC Curve and AUC Score?

The ROC Curve (Receiver Operating Characteristic) shows the trade-off between true positive rate and false positive rate across different thresholds. The AUC (Area Under the Curve) gives a single score to evaluate how well a model can distinguish between classes. AUC closer to 1 means better performance.

17. 🧮 What is the Curse of Dimensionality?

As the number of features or dimensions in a dataset increases, the data becomes sparse and difficult to interpret or visualize. This phenomenon is known as the Curse of Dimensionality.

Impacts:

  • Models become more complex and prone to overfitting
  • Computational cost increases
  • Distance metrics lose meaning

18. 🧬 What Are Confounding Variables?

Confounding variables are external factors that influence both the independent and dependent variables, creating a false association. In machine learning, failing to control for these can lead to misleading model predictions or biased results.

19. 🔄 What is the Difference Between Batch and Online Learning?

Batch Learning trains the model using the entire dataset at once. It's efficient for static datasets.
Online Learning updates the model incrementally as new data comes in—perfect for real-time or streaming data scenarios.

20. 🧱 What is One-Hot Encoding?

One-Hot Encoding is a technique used to convert categorical variables into a numerical format so they can be used in ML models. Each category becomes a new binary column (1 if present, 0 otherwise).

Example: Color = [Red, Blue, Green] becomes:

Red   Blue   Green
1      0      0
0      1      0
0      0      1

Comments

Post a Comment

Popular posts from this blog

Analysis of algorithms viva questions / Interview questions - set1 /sorting algorithms

Top 10 Questions on Sorting Algorithms and Divide and Conquer Approach Top 10 Questions on Sorting Algorithms and Divide and Conquer Approach Understanding basic sorting algorithms and the divide and conquer approach is crucial for every computer science student. Below are the most frequently asked questions with answers from Module 1: Introduction to Algorithms . 1. What is Selection Sort and how does it work? Selection Sort is a comparison-based algorithm that divides the array into a sorted and unsorted part. It repeatedly selects the minimum element from the unsorted section and moves it to the beginning. Time Complexity: O(n²) Stable: No In-place: Yes 2. What are the steps involved in Insertion Sort? Insertion Sort builds the sorted array one element at a time by inserting each new element into its correct position among the previously sorted elements. Best Case: O(n) Worst Case: O(n²) Stable: Ye...
Read more

Operating System Viva questions/interview questions 2025

1. What is an Operating System? An Operating System (OS) is system software that acts as a bridge between users and computer hardware. It manages hardware resources and provides services to application software. An OS handles tasks like memory management, process scheduling, and device control. It ensures that programs and users can run smoothly without interfering with each other. Examples include Windows, Linux, macOS, and Android. The OS also provides a user interface, either graphical (GUI) or command-line (CLI). Without an OS, a computer would be unusable. It organizes files, maintains security, and manages system performance. In short, the OS is the heart of any computing device. 2. What are the main functions of an Operating System? The primary functions of an Operating System are process management, memory management, file system management, and device management. It ensures efficient execution of processes by scheduling CPU time. Memory management keeps track of ea...
Read more

Recommendation System viva questions/ Interview questions

Question 1: Can you provide an overview of the history of recommendation systems and their evolution over time? Answer: Recommendation systems have evolved significantly since their inception in the early 1990s. Initially, simple systems such as collaborative filtering and content-based filtering emerged, followed by more sophisticated techniques like matrix factorization and deep learning. The rise of e-commerce and streaming platforms in the late 1990s and early 2000s led to increased research and development in recommendation algorithms. Today, recommendation systems are ubiquitous across various domains, from e-commerce and entertainment to social media and content platforms, and continue to evolve with advancements in machine learning and artificial intelligence. Question 2: What are the main functions of a recommender system, and how do they contribute to enhancing user experience and engagement? Answer: The main functions of a recommender syste...
Read more