NLP/TM - Providing Meaningful Insights from the Data

text_mining
r
nlp
data_science
python

#1

Hi,

I would like to know the possibility or feasibility of implementing the following;

Supposedly I have a Text Data, which consists two columns, namely Tweet and Time, for example:

[1] “its too noisy here in melbourne station”, 07:32 PM
[2] “whats up people, too much noise here is it”, 07:31 PM
and etc.,

I would like to have a output as - “After 7:30 PM, its too noisy in Melbourne Station”.

Is it possible to generate insights of these kinds using NLP? If so possible, any links and suggestions on implementing the same would be really helpful.

Thanks


#2

As per my knowledge haroon that requires some human intervention. You cannot completely automate that.

  1. I feel the semi automated way using reinforcement learning by training a model and assign rewards untill you get what you want.(Research area)

  2. Train a RNN for generating the insights if you have sufficient amount of data for each topic.

I recommend manual framing of sentence will be better if you have to provide business insights.