Thesis Idea Exploration

Начать. Это бесплатно
или регистрация c помощью Вашего email-адреса
Thesis Idea Exploration создатель Mind Map: Thesis Idea Exploration

1. Tools & Resources

1.1. NLTK Python Library for NLP Book has explanations and examples http://www.nltk.org/book/

1.2. COVID-19 Research Paper Dataset COVID-19 Open Research Dataset Challenge (CORD-19)

1.3. Python RAKE (2 versions) Rapid Automatic Keyword Extraction Called rake-nltk or python-rake

1.4. Tweepy Python library for using Twitter API

1.5. Applied NLP Course Notes by Berkeley Info 256. Applied Natural Language Processing

1.6. OpenNLP Apache OpenNLP is Java library Apache OpenNLP

1.7. CoreNLP Java, Stanford CoreNLP – Natural language software | Stanford CoreNLP

1.8. Gensim Python, supports large scale corpora, have parralel versions gensim: topic modelling for humans

1.9. spaCy

1.9.1. When to use which: Facts & Figures · spaCy Usage Documentation

2. Topics

2.1. Japanese

2.1.1. Improve tokenization Japanese lacks explicit word boundary markers (no spaces between words like in English)

2.2. COVID-19

2.2.1. Summarization of COVID Research Papers https://towardsdatascience.com/summarization-of-covid-research-papers-using-bart-model-5b109a6669a6

2.2.2. Summarization of Social Media Response to Government COVID-Related Actions

2.2.3. Twitter COVID Bot Detection/Analysis https://www.npr.org/sections/coronavirus-live-updates/2020/05/20/859814085/researchers-nearly-half-of-accounts-tweeting-about-coronavirus-are-likely-bots

2.2.4. Analysis of Emotional Response of Government COVID-Related Actions per Social Media Platform

2.2.5. Sentiment Analysis of Social Media Response to COVID-Related News/Gov Action

2.2.6. What connotations exist around different naming of the virus? (COVID-19, Coronavirus, etc.)

2.3. Consumer Products

2.3.1. Automatic summarization for consumer reaction to products via opinion mining

2.4. Hate Groups

2.4.1. Automate frame annotation through NLP/opinion mining Paper analyzed didn't automate much about the analysis of how different hate groups employ different social media platforms. All of it was manual, sociological analysis. Frame annotation = categorizing by diagnosis, motivation, etc.

2.5. Improve Abstractive Text Summarization

2.5.1. Primarily extractive-based techniques are used

3. Research Areas

3.1. Natural Language Processing

3.1.1. Opinion Mining/Sentiment Analysis

3.1.1.1. Overview of Opinion Mining http://ijiset.com/vol2/v2s6/IJISET_V2_I6_106.pdf High level. DOI: 10.1016/j.inffus.2016.10.004 In-depth.

3.1.2. Automatic Text Summarization

3.1.2.1. Overview of Extractive Techniques http://www.jcomputers.us/vol12/jcp1206-08.pdf https://arxiv.org/pdf/1707.02268.pdf

3.1.2.2. Subject, Predicate, Object tripes, remove nodes to create human-readable summary http://www.cs.cmu.edu/~jure/pubs/nlpspo-msrtr05.pdf

3.1.2.3. Opinion summarization by splitting microblog text segments into 8 emotion categories http://www.sciencedirect.com/science/article/pii/S0268401218313690 Found difficult to follow. Assumes only 3 emotional categories (pos, neg, neut).

3.1.2.4. Single document vs multi-document summarization

3.1.3. Language Correction

3.1.3.1. Japanese Error Correction https://www.aclweb.org/anthology/I11-1017.pdf Doesn't really follow my interests.

3.1.4. Text Classification

3.1.5. Chatbots, Speech Recognition, Automatic Text Generation, Question Answering, Subjectivity Production, Data Visualization

3.2. Social Computing

3.2.1. Hate Groups http://people.cs.vt.edu/tmitra/public/papers/hategroups-chi2020.pdf