Your task is to analyze an existing recommender system that you find interesting. You should:
Perform a Scenario Design analysis as described below. Consider whether it makes sense for your selected recommender system to perform scenario design twice, once for the organization (e.g. Amazon.com) and once for the organization’s customers.
Attempt to reverse engineer what you can about the site, from the site interface and any available information that you can find on the Internet or elsewhere.
Include specific recommendations about how to improve the site’s recommendation capabilities going forward.
Create your report using an R Markdown file, and create a discussion thread with a link to the Git Hub repo where your Markdown file notebook resides. You are not expected to need to write code for this discussion assignment.
Who are the target users?
The target users are all existing and prospective customers of Amazon.
What are their key goals?
The key goals are
How can we help them accomplish these goals?
The user should be able to quickly search for keywords for products on their mind. The user should be able to find recommendations of products that they might be interested in, or similar products that other users purchased when they were looking for that specific product.
Who are the target users?
The target users are the stakeholders like the shareholders, the management and the suppliers.
What are their key goals?
The key goal is to maximize long term shareholder growth by offering products that the customer is or might be interested in and transforming these leads into profits.
How can you help them accomplish these goals?
Match the customer needs with quality products and services, by recommending content that they might want to purchase.
Barnes and Nobles uses Collaborative Filtering (CF). CF uses the known preferences of a group of users to make predictions. CF can be memory based, model based or hybrid. Barnes and nobles uses the memory based CF.
Memory based CF can be user based or item based.
User based: Similar users who have similar ratings from similar items are found and then target user’s ratings are predicted for items they did not interact with.
For example, say Alice and Bob have given similar ratings to (or liked) some movies; Alice = {Terminator: 4, Predator: 2, Robocop: 3} Bob = {Terminator: 4, Predator: 2, Robocop: ?}
The steps in this technique are:
Item-based models find similar items to items which target user already rated or interacted.
Bob = {Matrix: 4, Kill Bill: 3, Terminator: 4} items Bob not rated= {Robocop, Alien}
The steps in this technique are:
I have always felt that the customer review section could have a chatbot to answer specific questions about the product. There could also be a portal where you may discuss with experts (especially when it comes to highly technical books).
The other lack I have felt is the myriad of recommendations. Sometimes I click on products just based on curiousity but not really interested to purchase the product. I believe customers should be able to choose core areas that they are actually interested in and have the choice to limit email based recommendations based on their choices.