Natural Language Processing

Singapore University of Social Sciences

Natural Language Processing (AIB551)

Synopsis

This course AIB551 Natural Language Processing provides an introduction to novel and effective machine learning techniques to understand and analyse human language for business applications. The module will explain the key approaches of acquiring and handling natural textual information using Python. Students will learn to pre-process and transform unstructured/semi-structured textual data into analysable format. Next, students will learn advanced data science solutions created for Natural Language Processing (NLP) including word, sentence and document embeddings, word clouds, Named Entity Recognition (NER), sentiment analysis, topic modelling, and influencer/topic network analysis enabled by supervised, unsupervised and deep machine learning techniques. Real-world examples, hands-on class exercises and assignments will be designed to help students to prepare and analyse data by harnessing Python’s libraries and using NLP techniques to extract new knowledge being integrated into data analysis workflows and improve business performance.

Level: 5
Credit Units: 5
Presentation Pattern: EVERY REGULAR SEMESTER

Topics

Basics of Natural Language Processing (NLP)
Build NLP Vocabulary Using Python
Bag-of-Words (BoW) and TF-IDF Vectors
Pre-trained Word2vec Model
Venture into Doc2vec, Sent2Vec and Universal Sentence Encoder
Word Cloud Analysis
Standard and Customised Sentiment Analyser
Entity Recognition Techniques
Evaluation Metrics of Classifier
Topic Modelling Techniques
Topic and Influencer Network
State of the Art in NLP - Artificial Neural Networks (ANNs), Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs)

Learning Outcome

Develop know-how of the applications of AI tools and technologies in areas of NLP
Assess the possibilities and implications of NLP in different industries
Formulate NLP analytics strategies
Prepare unstructured/semi-structured data into analysable format
Design NLP solutions through supervised, unsupervised and deep machine learning techniques
Revise data analysis workflows by integrating NLP solutions to improve business performance