Repository for AIForBusiness NLP Use Cases
This is Hands On Exercise to processing Textual Data for Insights using NLP techniques.
DataSet : Here we use Airline Tweets or Textual Comments / Statements made by Travellers / users. Objective : Our Objective is to classify tweets as positive or negative so that suitable business action can be taken based on classification.
Dataset Details : This dataset is as provided by Kaggle. The dataset has the following columns:
tweet_id
airline_sentiment
airline_sentiment_confidence
negativereason
negativereason_confidence
airline
airline_sentiment_gold
name
negativereason_gold
retweet_count
text
tweet_coord
tweet_created
tweet_location
user_timezone
Few many Attribtues (or Columns or Features) of interest could be : airline_sentiment and text.
Hands-on Exercises to learn NLP
a) Lesson 01 NLP Warm-up: In this first Hands-on Exercise students will be able to work on various Text Processing concepts as it relates to NLP:
- Reading Text documents
- Manipulating Strings
- NLTK Library
- Corpus, Language, Vocabulary
- Tokenization
- Regular Expressions and Patterns
- Stop Words processing
- Stemming
- Lemmatization
- Parts of Speech
NLP Pipeline Processing : Walkthrough
b) Lesson 02 End to End NLP Example: Airline Sentiment Classification Case Study. In this second lesson we will use a proper dataset to built a Classfication model on the back of NLP based text data processing.
- Data Acquistion
- Data Processing
- Data Preparation using NLP concepts like Stop Words, Stemming etc.
- Vectorization, Embedding
- Train and Test split.
- Build a Classification Model
- Model Inference
- Evaluation and Application