Analyze the theoretical foundation of machine learning to determine how an the intelligent machine works.
LO2 Investigate the most popular and efficient machine learning algorithms used in LO2
industry
LO3 Develop a machine learning application using an appropriate programming
language or machine learning tool for solving a real-world problem
LO4 Evaluate the outcome or the outcome of the application to determine the
effectiveness of the learning algorithm used in the application.
Transferable skills and competencies developed
Machine learning concepts, machine learning algorithms, gain hands-on experience in implementing algorithms using a programming language
Vocational scenario:
You're working as a Machine Learning Engineer at the Amazon company. Your role consists of developing innovative solutions using machine learning to solve problems and create valuable solutions for the company's customers.
A part of your role is discussing with the team, choosing suitable implementation strategies and algorithmic decisions, as well as documentation.
Example scenarios on where the company needs machine learning implementation:
Amazon provides Dashboards for their customers to show meetings, time, weather and personalized features, including intelligent features for weather prediction and stock price prediction.
- Amazon provides WBS which is cloud infrastructure for various clients and agencies, Get Your Car (GYC) is a client for Amazon and hold all of their data on WBS, they need some predictions from Amazon on their cloud hosting, for example, according to the users data, they need to tell whether a user will own a car or not.
Amazon needs to be able to distinguish spam reviews from realistic (non-spam) reviews on their main shopping website, according to certain details such as: account name, creation date, regular behavior, etc.

Based on these scenarios, you will test and implement machine learning algorithms to provide the support that amazon needs, follow the tasks below to prepare all deliverables of this project.

Consider the following datasets for your experiments, and if you find interesting results, you will use the experimental algorithms on Amazon platforms:
- Testing Dataset 1: Weather Dataset: Predicting the temperature, based on Humidity.
- Testing Dataset 2: Car Ownership Dataset: Indication whether an individual owns a car or not, according to their personal data.
- Testing Dataset 3: Simulating the future value of investments in Amazon, according to interest rates going up and down through the years.
- Dataset 4: Amazon Datasets (choose any of the datasets in the list): https://www.kaggle.com/datasets/naveedhn/amazon-product-review-spam-and-non-spam

The questions are:

 Assignment activity and guidance

Task 1: Discussing the company needs

 

1. Present and discuss with your team, how every problem mentioned above, is different than the other problems, what characterizes every problem?
2. According to you discussion, outline the type of every machine learning problem you are facing, and how would you describe every category of machine learning problems?
3. How would Amazon detect spam reviews from customers over the products? Using machine learning algorithm vs. A non machine learning algorihtm, evaluate this case.
4. Alexa for example, is one of Amazon's intelligent systems,Critically Assess the reasons why Amazon need machine learning algorithms to be implemented when desinging Alexa.

 

Task 2: Testing and Investigating different development options

1. Examin the use of Decision Tree, Random forest, naïve bayes, classifiers over test dataset 1, and Examin K-means algorithm over test dataset 2 and Monte Carlo algorithm over test dataset 3 Show how would these algorithms help with getting good results over these datasets for testing purposes.
2. Present your results for every algorithm, to prove their efficency to Amazon engineers, and to be able to compare their performance.
3. Take Case 1 dataset, Explore the strength points in the algorithms mentioned in point 5 according to the performance over the testing dataset that you have.

 

Task 3: Solution Implementation

1. Use the Amazon dataset attached with this assignment, Split the data set into training and testing, and explain in a small paragraph how should the data be prepared before implementing the machine learning model.
2. Based on the algorithms you investigated in task 2, Choose any algorithm you find suitable, to solve Amazon's problem of classifying user reviews to spam or not spam.
3. Build the machine learning model based on the algorithm you chose, over the data provided, show the model results including (accuracy, AUC, F-measure, confusion matrix).
4. Explain to you colleages every step of building this model, focus on testing results, your explanations should be as markdown boxes inside your jupyter notebook.
5. Based on your testing results critically asses how your machine learning model is effective for the mentioned scenario and data and useful for amazon, explain how effective it is to meet Amazon's goal in this project?
6. Consider whether the output of the algorithm you have picked previously has the problems of (underfitting or over-fitting), provide explanation according to your results. Also, based on your results, examin the effectiveness of your model, do you think the algorithm you chose is suitable here according to this problem? Assess the effectiveness of this algorithm in general.