4

I am trying to figure out if the word recommend is the word I am actually looking for.

Background

I am trying to decide whether determine if an algorithm I am writing is a Recommender System or something else. In my case my algorithm learns from the history of the user and picks (this is important as the user does not pick, the algorithm forcefully picks) a certain strategy after that history. This is why I'm thinking the term recommend is wrong as the user does not have a chance to choose.

The algorithm spits out a list of things and it's probability of it being in the users favor.

option A 0.32
option B 0.1121
option C 0.054
etc.

So I guess it does recommend the list in this order, but then I choose the top option A always as.

So I think of my system in two ways.

  • It IS a Recommender System, just that I choose the option with the largest probability.

  • Well, it is NOT, it is something else.

Meaning

Meanings of the word from top results on Google:

put forward (someone or something) with approval as being suitable for a particular purpose or role.

to say that (someone or something) is good and deserves to be chosen

to present as worthy of confidence, acceptance, use, etc.; commend; mention favorably:

Also the issue I have with the word is that the meaning of recommend to me is normally used when an object is being recommended is not very known to the party in which the recommendation is being directed at.

I am trying to find a word in which the party already knows of it.

Other words

If there are any other words which can be suitable by all means please tell me. In the Computer Science arena the following terms pop up when in the context of learning from the user and doing something with that data:

Personalization Context-awareness Adaptive personalization

2
  • Full points for realizing that the users don't have a choice. However, you should still use recommend as this is the standard word for the situation. A lot of websites track your history and suggest new things based on that. They usually call them recommendations.
    – Tushar Raj
    Commented May 15, 2015 at 21:35
  • Personalization is slightly different than recommendation. It's not limited to suggestions. It can also mean enhancing the overall experience of the user (look and feel, custom feeds etc). Also, it's quite normal for the user to have a say in the process.
    – Tushar Raj
    Commented May 15, 2015 at 21:40

2 Answers 2

1

I would naively call it a 'Recommendation Heuristic" because it is using a ruleset to make an educated guess with existing data (data which is growing over time) the aim being to make a satisfactory solution (optimal being out of reach because of incomplete data) - It is a heuristic that generates recommendations!

1

Whenever you're talking about naming an algorithm, I think the more detail you can provide the better. Ultimately, the quality of the name will depend on how natural the analogy translates from a familiar "real-world" concept to the "binary"/"digital" concept you're inventing within your application domain.

Given your description, I would call the portion of the software that produces the recommendations (i.e. options A, B, C, and their corresponding ratings) something akin to "Recommendation Engine/System/Service".

As for the portion that drives the user through the application based on those recommendations, a lot of that will depend on the application domain itself. Whatever most closely matches "making choices on behalf of the user to drive/steer/direct/etc. them through the application". Some things that come to mind for that "decider" element:

  1. Routing System
  2. Navigator System
  3. Autopilot System
  4. Ideation Engine
3
  • Updated question! Commented May 15, 2015 at 22:17
  • As stated though, the user does not have a choice at all in this. This is where my question comes in! Commented May 15, 2015 at 22:18
  • Updated the answer, sorry I went MIA there on you--was away from a computer for a couple days here. Does my update help?
    – AGB
    Commented May 16, 2015 at 22:52

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .