Question 1

What is Pre-Diabetes?

Accepted Answer

Pre-diabetes mellitus (Pre-DM) is when blood-sugar levels are higher than normal; warning sign that someone is at risk of developing type2 diabetes, heart disease or stroke. 
'Mellitus', latin term that means 'sweet'

Question 2

What is Supervised Learning?

Accepted Answer

The process of providing an algorithm with records in which an output variable of interest is known and the algorithm “learns” how to predict this value with new records where the output is unknown.

It’s like training the machine with examples that already have the correct answers, so it can learn to make predictions on new data.

Question 3

What is the typical life cycle of a data science project?

Accepted Answer

1. Problem Definition – Define the question and goals.
2. Data Collection – Gather data from relevant sources.
3. Data Cleaning – Handle missing values, outliers, and inconsistencies.
4. Exploratory Data Analysis (EDA) – Visualize and summarize data patterns.
5. Feature Engineering – Create and select useful features.
6. Model Building – Train models using appropriate algorithms.
7. Model Evaluation – Assess performance with metrics and validation.
8. Deployment – Integrate the model into a real-world system.
Monitoring – Track performance and update as needed.

Question 4

What is data preprocessing in a data science workflow?

Accepted Answer

The step of data preprocessing involves verifying data quality, handling missing values, addressing outliers in the data, and ensuring consistency.

Question 5

What is the Logistic Regression model in machine learning and data science?

Accepted Answer

Logistic Regression is a statistical method used for binary classification—predicting one of two possible outcomes (e.g., yes/no, spam/not spam). This method models the probability that a given input belongs to a particular class using the logistic (sigmoid) function.

To map Z to a probability between 0 and 1, logistic regression uses the following function:
             P (y = 1 | X) = σ(Z) = 1/1+e^-z
This gives the probability that the output y is 1 (positive class), given the input X.

Question 6

What is the Support Vector Machine model in machine learning and data science?

Accepted Answer

Support Vector Machine (SVM) is a powerful supervised machine learning algorithm used for classification and regression tasks as well as outliers detection. It is highly effective in cases where the number of dimensions is greater than the number of samples. The SVM algorithm sorts data by finding the best line that separates different classes, making sure there’s as much space as possible between them.

Question 7

What is the Random Forest model in machine learning and data science?

Accepted Answer

Random Forest is a machine learning method used to make predictions. It combines the results of multiple decision trees to make a more accurate prediction. Decision trees can be visualized as a flow of chart of conditions to most accurately classify or predict outcomes. It results in a higher accuracy, avoids overfitting and works well with many data types.

Question 8

What is the XGBoost model in machine learning and data science?

Accepted Answer

XGBoost (eXtreme Gradient Boosting) is a scalable, gradient boosted decision tree machine learning library mainly used for classification and regression problems. It is known for being fast, accurate and efficient. It works well with both large and small datasets, handles missing data and uses techniques such as regularization to avoid overfitting.

Question 9

What is cross-validation in machine learning and why is it important?

Accepted Answer

Cross-validation is a way to check how well a machine learning model will work on new data. Instead of just testing the model once, we test it multiple times using different parts of the data.

The cross-validation method tests the model several times using different combinations of the data, not just one, to make sure it works well no matter how the data is split. Additionally, it helps ensure the model is reliable, works well on unseen data, and is especially useful when data is limited.

Question 10

What is overfitting in a machine learning model?

Accepted Answer

Overfitting is where a model is too closely matched to the training data, so it not only learns the important patterns but also picks up on random details that aren’t really useful. This makes the model work well on the training data but poorly on new data.

Question 11

What is LASSO (L1) and Ridge (L2) regularization in data science?

Accepted Answer

In the field of data science, models such as linear regression are built to make predictions. But, a model may fit the training data too well. This is called overfitting. 
Regularization helps prevent overfitting by adding a penalty to the model for being too complex.

LASSO (L1) Regularization adds a penalty based on the absolute values of the coefficients. Essentially, it tries to simplify the model by choosing only the most important features and ignoring the rest.
Mathematical approach:
λ * sum of (absolute values of coefficients)

Ridge (L2) Regularization adds a penalty based on the squares of the coefficients. Essentially, it uses all the features, but keeps their influence small. 
Mathematical approach:
λ * sum of (coefficients²)

Question 12

What is the purpose of data visualization in data science and how does it enhance data interpretation?

Accepted Answer

Data visualization helps us see patterns, trends, and important information in the data by turning it into graphs or charts. It makes it easier to understand what the data is telling us, which helps in making better decisions and communicating findings clearly.

Question 13

What are SHAP values in data visualization?

Accepted Answer

SHAP (SHapley Additive exPlanations) values show how much each feature contributes to a model’s prediction by indicating how it shifts the output from the average. SHAP plots improve interpretability, emphasize important features, and help build trust in complex models like XGBoost by making their decisions more transparent.

Question 14

What is a pairplot and how is it used in exploratory data analysis?

Accepted Answer

A pairplot is a type of graph that helps us see how different features (columns) in a dataset are related to each other. It creates a grid of scatter plots, where each plot compares two features at a time. This way, we can quickly understand how features change together. In the process of exploratory data analysis, a pairplot helps us see relationships between features, spot patterns, identify outliers, and understand how features behave both individually and in relation to each other.

Question 15

What is a confusion matrix and what is its purpose in evaluating classification models?

Accepted Answer

A confusion matrix is a table that shows how well a model’s predictions match the actual results. It compares the predicted values (what the model thought) with the true values (what the actual answer should be).

True Positive (TP): The model correctly predicted a positive result (e.g., spam email marked as spam).
True Negative (TN): The model correctly predicted a negative result (e.g., not spam email marked as not spam).
False Positive (FP): The model wrongly predicted a positive result (e.g., not spam email marked as spam).
False Negative (FN): The model wrongly predicted a negative result (e.g., spam email marked as not spam).

A confusion matrix gives a clear view of where the model is correct and where it needs improvement.

Bhavana Kannan

First-year student studying Computer Science (Software Engineering) at ASU

Explore other Linktrees

About this account

More from Linktree