Machine Learning Pipeline

Of the following, which is an example of machine learning? (Select TWO.)

• A. Calculating the shortest route from current location to the destination

• B. Optimizing product pricing based on real-time sales data

• C. Sentiment analysis of text on product reviews

• D. A loan approval system that classifies applicants entirely based on credit score

Which of the following is an appropriate use case for unsupervised learning?

• A. Partitioning an image of a street scene into multiple segments

• B. Finding an optimal path out of a maze

• C. Identifying clusters of housing sales based on related data points

• D. Analyzing sentiment of social media posts

A retail company wants to deploy a machine learning model to predict the demand for a product using sales data from the past 5 years.

What is the MOST efficient solution that the company should implement first?

• A. Regression

• B. Multi-class classification

• C. Binary class classification

In which phase of the ML pipeline do you analyze the business requirements and re-frame that information into a machine learning context.

• A. Problem formulation

• B. Model training

• C. Deployment

• D. Data preprocessing

Which feature of Amazon SageMaker can you use to learn patterns in data?

• A. SageMaker notebook instances

• B. SageMaker training jobs

• C. SageMaker hyperaparameter tuning

• D. SageMaker endpoints

A retail company wants to start personalizing product recommendations to visitors of their website. They have

2 historical data of what products the users have purchased and want to implement the system for new users, prior to them purchasing a product. What's one way of phrasing a machine learning problem for this situation?

• A. Predict if a user will buy the product based on all the other products they bought

• B. Classify the products to categories to recommend the user based on the most common category bought

• C. Predict the next item that user will buy based on every the products all users bought

• D. Predict the item the user will buy based on other users' purchase history and the product ratings and reviews

A Data Scientist at a retail company is using Amazon SageMaker to classify social media posts that mention the company into one of two categories: Posts that require a response from the company, and posts that do not. The Data Scientist is using a training dataset of 10,000 posts, each of which contain the timestamp, author, and full text of each post. However, the Data Scientist is missing the target labels that are required for training.

Which approach can the Data Scientist take to create valid target label data? (Select TWO.)

• A. Ask the social media handling team to review each post using Amazon SageMaker GroundTruth and provide the label

• B. Use the sentiment analysis natural language processing library to determine whether a post requires a response

• C. Use Amazon Mechanical Turk to publish Human Intelligence Tasks that ask Turk workers to label the posts

• D. Use the a priori probability distribution of the two classes. Then, use Monte-Carlo simulation to generate the labels

• E. Use K-Means to cluster posts into various groups, and pick the most frequent word in each group as its label

An ad tech company is using an XGBoost model to classify its clickstream data. The company’s Data Scientist is asked to explain how the model works to a group of non-technical colleagues.

What is a simple explanation the Data Scientist can provide?

• A. XGBoost is an Extreme Gradient Boosting algorithm that is optimized for boosted decision trees

• B. XGBoost is a logistic regression algorithm to split each feature of the data and used for classification problem

• C. XGBoost is a robust, flexible, scalable algorithm that uses linear regression and used for regression problems

• D. XGBoost is an efficient and scalable neural network architecture

You are preprocessing a credit card fraud dataset that contains the column 'Amount'. You want to determine what the average amount is for the cases where there is fraud. Which basic descriptive statistic could you use to achieve this?

• A. Mode

• B. Median

• C. Standard Deviation

• D. Mean

An ML Engineer at a real estate startup wants to use a new quantitative feature for an existing ML model that predicts housing prices. Before adding the feature to the cleaned dataset, the Engineer wants to visualize the feature in order to check for outliers of the feature.

What visualization technique should the ML Engineer use?

• A. Box Plot

• B. Histogram

• C. Scatterplot

• D. Heatmap