Skip to content

Latest commit

 

History

History
28 lines (19 loc) · 903 Bytes

File metadata and controls

28 lines (19 loc) · 903 Bytes

Api

The model used for api is LabelPowerset Transformation with Random Forest Classifier

Route Information :

Api route : apiforprogrepper.herokuapp.com/api/companies

Rquest Type : POST

Params :

  • problem-statement: ' ' // problem statement text
  • tags : [ ] //list of tags default empty
  • num-companies: 5 //top n most probabilistic companies

Flow :

  1. Data received by the api is passed through the following functions :
  • HTML tags are removed if any
  • Punctuation symbols are removed if any
  • Other than alphabets, other symbols are removed
  • Stop words are removed
  • Stemming is done to convert the tokens into their base form
  1. The preprocessed data is then fit into the vectorizer (vectorizer_new.sav)

  2. Finally,the data is passed through the model(company_tags_lbps_lg_new.sav) and the probabilities are calculated.

  3. The top n probabilistic companies are returned.