2

My project needs some natural language processing. I'm completely new to the field.

what I'm trying to achieve is that when the User enter the description of the product I look for in my database which description is nearest and suggest that the category, product group and sub-group (the tree of the product).

For this titles 250 extracts products for each subgroup.

What is the specific term in NLP for doing this? I tried googling for a while, but had no luck since I don't know the term. Any good tutorials to start with? Are there any good libraries in doing this specific task?

Thank you.

sn3fru
  • 106
  • 3
  • 12

1 Answers1

3

From what I can tell autocomplete or text prediction/predictive search isn't really a big research area in NLP. It wasn't even covered in any of my graduate level classes and I do research in this area. I think the reason is that there are solutions that exist which are good enough for the vast majority of real world problems.

I'm not sure which language you work in, but the library you want to work with is probably Lucene if you are dealing with java, perhaps setting up a Solr instance if this is a general problem for you and you are dealing with a large number of ontologies.

You can find some reason tutorials/examples here on stack overflow, such as: How to implements auto suggest using Lucene's new AnalyzingInfixSuggester API?

Community
  • 1
  • 1
ozborn
  • 980
  • 5
  • 24
  • Many thanks for the reply. I will try to make my question more directly. I am creating a website ads and I find the product tree (category, group and sub-group) the published announcement. For example, the User wants to sell an iphone and colleague the title "iphone 6 white with 32gb." I want to identify that the category is telephony, the group is mobile and the subgroup is iphone. For I created all possible tree and have 250 ad titles of examples in each subgroup. – sn3fru Jan 14 '16 at 00:29
  • When the User type ~ "iphone 6 white with 32gb" it is possible to appear many subgroups such as "accessories", "computer", etc.I need suggest that best matches, or category, the group and subgroup correct. – sn3fru Jan 14 '16 at 00:29