NER or Text extraction in NLP?


#1

Hi,
I have a problem of Text extraction. like suppose below are two paragraph are part of 2 different company profile and in that I need to find company business segment.

  1. ASC 280-10 defines operating segments as components of a company about which separate financial information is available that is evaluated regularly by the chief decision maker in deciding how to allocate resources and in assessing performance. The Company has two segments: App development and Training.
    2.The Company’s chief operating decision-maker is considered to be the Chief Executive Officer (“CEO”). The CEO reviews financial information for purposes of making operational decisions and assessing financial performance. This financial information is consistent with the information presented in the accompanying statements of operations. The Company operates in one reportable segment, the education market.

I made the segment in bold letter in above 2 paragraph.
This does not comes under traditional NER, as the entity does not comes under any unique grammar, so not able to do chunking.

Any idea to solve this. If any one wants more details please reply, will elaborate more.


#2

Hi @tell2jyoti

Given you have plenty of labeled data, you can do this by following:

  1. Cleaning the text data by removing the stopwords, punctuations, etc (TF-IDF might help as well).

  2. Training an algorithm for mapping each observation/ comment with the respective class/ (market segment).

  3. Predicting for the new observations.

You will find this resource helpful:

Best,
Saurav.