## Blog

# Top 10 Interview Question of Machine Learning

- May 18, 2023
- Posted by: jcodebook
- Category: Machine Learning

**1. W๐ต๐ฎ๐ ๐ถ๐ ๐๐ต๐ฒ ๐ฑ๐ถ๐ณ๐ณ๐ฒ๐ฟ๐ฒ๐ป๐ฐ๐ฒ ๐ฏ๐ฒ๐๐๐ฒ๐ฒ๐ป ๐๐๐ฝ๐ฒ๐ฟ๐๐ถ๐๐ฒ๐ฑ ๐ฎ๐ป๐ฑ an ๐๐๐ฝ๐ฒ๐ฟ๐๐ถ๐๐ฒ๐ฑ ๐น๐ฒ๐ฎ๐ฟ๐ป๐ถ๐ป๐ด?**

Supervised learning involves training a model on labeled data to make predictions, while unsupervised learning involves finding patterns, relationships, and structures in unlabeled data without specific target outcomes.

Labeled data refers to sets of data that are given tags or labels, and thus made more meaningful.

**2.๐ช๐ต๐ฎ๐ ๐ถ๐ ๐๐ต๐ฒ ๐๐ฒ๐ป๐๐ฟ๐ฎ๐น ๐๐ถ๐บ๐ถ๐ ๐ง๐ต๐ฒ๐ผ๐ฟ๐ฒ๐บ (CLT) ๐ฎ๐ป๐ฑ ๐๐ต๐ ๐ถ๐ ๐ถ๐ ๐ถ๐บ๐ฝ๐ผ๐ฟ๐๐ฎ๐ป๐?**

The Central Limit Theorem (CLT) states that the distribution of sample means approximates a normal distribution, even if the population isnโt normally distributed. It is crucial in inferential statistics as it allows us to make inferences about a population based on a sample.

A normal distribution is a symmetrical, bell-shaped distribution with no skew.

**3.E๐
๐ฝ๐น๐ฎ๐ถ๐ป ๐๐ต๐ฒ ๐ฐ๐ผ๐ป๐ฐ๐ฒ๐ฝ๐ ๐ผ๐ณ ๐ฟ๐ฒ๐ด๐๐น๐ฎ๐ฟ๐ถ๐๐ฎ๐๐ถ๐ผ๐ป ๐ถ๐ป ๐บ๐ฎ๐ฐ๐ต๐ถ๐ป๐ฒ ๐น๐ฒ๐ฎ๐ฟ๐ป๐ถ๐ป๐ด.**

Regularization is a technique used to prevent overfitting in machine learning models. It adds a penalty term to the loss function, discouraging complex models and promoting simplicity, ultimately improving the model’s generalization performance.

**4.๐ช๐ต๐ฎ๐ ๐ถ๐ ๐ณ๐ฒ๐ฎ๐๐๐ฟ๐ฒ ๐๐ฒ๐น๐ฒ๐ฐ๐๐ถ๐ผ๐ป ๐ฎ๐ป๐ฑ ๐๐ต๐ ๐ถ๐ ๐ถ๐ ๐ถ๐บ๐ฝ๐ผ๐ฟ๐๐ฎ๐ป๐?**

Feature selection is the process of selecting the most relevant and informative features from a dataset. It helps improve model performance by reducing dimensionality, eliminating noise, and enhancing interpretability.

**5.๐ช๐ต๐ฎ๐ ๐ถ๐ ๐๐ต๐ฒ ๐ฑ๐ถ๐ณ๐ณ๐ฒ๐ฟ๐ฒ๐ป๐ฐ๐ฒ ๐ฏ๐ฒ๐๐๐ฒ๐ฒ๐ป ๐ฏ๐ฎ๐ด๐ด๐ถ๐ป๐ด ๐ฎ๐ป๐ฑ ๐ฏ๐ผ๐ผ๐๐๐ถ๐ป๐ด?**

Bagging and boosting are ensemble learning techniques. Bagging involves training multiple models independently on different subsets of the training data and aggregating their predictions, while boosting focuses on sequentially training models, giving more weight to misclassified instances to improve overall performance.

**6.๐ช๐ต๐ฎ๐ ๐ถ๐ ๐ฐ๐ฟ๐ผ๐๐-๐๐ฎ๐น๐ถ๐ฑ๐ฎ๐๐ถ๐ผ๐ป ๐ฎ๐ป๐ฑ ๐๐ต๐ ๐ถ๐ ๐ถ๐ ๐๐๐ฒ๐ณ๐๐น?**

Cross-validation is a technique used to evaluate a model’s performance by partitioning the data into multiple subsets. It helps assess the model’s ability to generalize to unseen data and aids in hyperparameter tuning and model selection.

**7.๐๐
๐ฝ๐น๐ฎ๐ถ๐ป ๐๐ต๐ฒ ๐ฏ๐ถ๐ฎ๐-๐๐ฎ๐ฟ๐ถ๐ฎ๐ป๐ฐ๐ฒ ๐๐ฟ๐ฎ๐ฑ๐ฒ๐ผ๐ณ๐ณ.**

The bias-variance tradeoff refers to the balance between a model’s ability to accurately capture the true underlying patterns (low bias) and its sensitivity to fluctuations in the training data (low variance). Finding the optimal tradeoff is crucial to avoid underfitting or overfitting.

**8.๐ช๐ต๐ฎ๐ ๐ถ๐ ๐/๐ ๐๐ฒ๐๐๐ถ๐ป๐ด ๐ฎ๐ป๐ฑ ๐ต๐ผ๐ ๐ถ๐ ๐ถ๐ ๐๐๐ฒ๐ฑ ๐ถ๐ป ๐ฑ๐ฎ๐๐ฎ ๐๐ฐ๐ถ๐ฒ๐ป๐ฐ๐ฒ?**

A/B testing is a statistical technique used to compare two versions of a variable (A and B) to determine which performs better. In data science, it is commonly used to evaluate the impact of changes in a product or system, helping in decision-making and optimization.

**9.๐๐ผ๐ ๐ฑ๐ผ ๐๐ผ๐ ๐ต๐ฎ๐ป๐ฑ๐น๐ฒ ๐บ๐ถ๐๐๐ถ๐ป๐ด ๐๐ฎ๐น๐๐ฒ๐ ๐ถ๐ป ๐ฎ ๐ฑ๐ฎ๐๐ฎ๐๐ฒ๐?**

Missing values can be handled through techniques such as imputation (replacing missing values with estimated values), deletion (removing instances with missing values), or using algorithms specifically designed to handle missing data.

**10.๐ช๐ต๐ฎ๐ ๐ถ๐ ๐๐ต๐ฒ ๐ฑ๐ถ๐ณ๐ณ๐ฒ๐ฟ๐ฒ๐ป๐ฐ๐ฒ ๐ฏ๐ฒ๐๐๐ฒ๐ฒ๐ป ๐ผ๐๐ฒ๐ฟ๐ณ๐ถ๐๐๐ถ๐ป๐ด ๐ฎ๐ป๐ฑ ๐๐ป๐ฑ๐ฒ๐ฟ๐ณ๐ถ๐๐๐ถ๐ป๐ด?**

Overfitting occurs when a model performs well on the training data but poorly on unseen data due to capturing noise or irrelevant patterns. Underfitting happens when a model is too simple and fails to capture the underlying patterns in the data.