You work as a machine learning specialist for a political candidate that is mounting a campaign to get reelected in her US senate district.
Your job is to build a machine learning model that allows the campaign to understand how to reach groups of similar counties by highlighting messages that resonate with those groups.
The senate candidate has a limited budget.
So, you need to build a cost-effective solution.
Which machine learning services and features should you use to solve this problem?
Click on the arrows to vote for the correct answer
A. B. C. D.Correct Answer: B.
Option A is incorrect.
Using Kinesis Data Streams and writing a Kinesis Data Streams Client Library application will be more costly than using Kinesis Data Firehose and its Lambda blueprint capability.
Also, the Factorization Machines algorithm is not used for clustering groups as this scenario requires.
Option B is correct.
Kinesis Data Firehose and its Lambda blueprints capability allow you to create the data gathering part of your machine learning solution at a lower cost than the other options.
Also, the K-Means algorithm is the best algorithm to use for clustering of groups, as this scenario requires.
Option C is incorrect.
You could use a Glue ETL job to gather the census data from the API.
However, the K-Nearest Neighbors algorithm is not a good choice for the clustering of groups as this scenario requires.
Option D is incorrect.
Kinesis Data Firehose and its Lambda blueprints capability allow you to create the data gathering part of your machine learning solution at a lower cost than the other options.
However, the K-Nearest Neighbors algorithm is not a good choice for the clustering of groups as this scenario requires.
References:
Please see the Amazon SageMaker developer guide titled K-Means Algorithm (https://docs.aws.amazon.com/sagemaker/latest/dg/k-means.html),
The Amazon SageMaker developer guide titled K-Nearest Neighbors (k-NN) Algorithm (https://docs.aws.amazon.com/sagemaker/latest/dg/k-nearest-neighbors.html),
Amazon SageMaker Examples GitHub repository titled Analyze US census data for population segmentation using Amazon SageMaker (https://github.com/aws/amazon-sagemaker-examples/blob/master/introduction_to_applying_machine_learning/US-census_population_segmentation_PCA_Kmeans/sagemaker-countycensusclustering.ipynb),
The Amazon SageMaker developer guide titled Factorization Machines Algorithm (https://docs.aws.amazon.com/sagemaker/latest/dg/fact-machines.html),
The Amazon Kinesis Data Firehose developer guide titled Amazon Kinesis Data Firehose Data Transformation (https://docs.aws.amazon.com/firehose/latest/dev/data-transformation.html)
The best solution for this scenario is option D.
Explanation: The first step in the process is to gather US anonymized census data from the US census on demographics by different US counties using the US Census Bureau Data API. The data should then be streamed to Kinesis Data Firehose. Kinesis Data Firehose is a cost-effective way to ingest streaming data, and it can easily deliver the data to various destinations, including Amazon S3.
Once the data is ingested into Kinesis Data Firehose, the next step is to perform feature engineering on the data. Feature engineering is the process of selecting and transforming variables in a dataset to improve the performance of machine learning models. In this scenario, feature engineering is necessary to identify messages that resonate with similar groups of counties.
To perform feature engineering, we can use the Kinesis Data Firehose Lambda blueprints to create a Lambda function. This Lambda function will be used as transformations to perform feature engineering on the data and write it to S3. The Kinesis Data Firehose Lambda blueprints are a set of pre-built templates that can be used to create Lambda functions. They can be customized to fit specific use cases, making it a cost-effective way to perform data transformations.
Finally, we can use the K-Nearest Neighbors (KNN) SageMaker built-in algorithm to produce the similar counties analysis to be used in the advertising for the grouped counties. The KNN algorithm is a simple and effective way to identify similar groups of counties based on their demographic data. It can be used to create a recommendation engine that suggests messages that resonate with similar groups of counties.
Overall, option D is the best solution for this scenario because it leverages cost-effective services such as Kinesis Data Firehose and the Kinesis Data Firehose Lambda blueprints to perform data transformations. Additionally, the KNN algorithm is a simple and effective way to identify similar groups of counties based on their demographic data.