Page 252 - Ai_C10_Flipbook
P. 252
NLP with Orange Data Mining Tool
NLP with the Orange Data Mining Tool allows users to perform text preprocessing, topic modelling, and sentiment
analysis through an easy-to-use visual workflow interface. It simplifies natural language processing tasks without
requiring extensive programming knowledge.
Sentiment Analysis Methods
Sentiment Analysis Methods can be used to determine the sentiment expressed in a piece of text, such as a
review, comment, or social media post. It helps to identify whether the expressed sentiment is positive, negative,
or neutral. Some of the Sentiment Analysis methods are as follows:
• Liu Hu: It uses English and Slovenian languages. It utilises a lexicon-based approach for sentiment analysis. This
means it relies on predefined dictionaries of positive and negative words to determine sentiment.
• Vader (Valence Aware Dictionary and Sentiment Reasoner): It uses English language. It is a lexicon- and
rule-based method designed to capture both the intensity (valence) and polarity (positive, negative, neutral) of
sentiments in text.
• Multilingual Sentiment: It supports multiple languages. It is also lexicon-based, but focuses on accommodating
several languages for broader applicability.
• SentiArt: It uses English and German. It is based on vector space models, which are computational methods
representing words or phrases in multi-dimensional space, it computes the valence (emotional value) of text.
• Custom Dictionary: It allows users to define dictionaries with their own positive and negative terms. It offers
flexibility by enabling users to tailor sentiment dictionaries to specific contexts. The analysis follows Liu Hu's
computation method.
Reboot
1. What NLP tasks can be performed using the Orange Data Mining Tool?
2. What languages does SentiArt support for sentiment analysis?
CASE STUDY: Sentiment Analysis For Internet Movie Database (IMDb)
IMDb—The Internet Movie Database (IMDb) is an online database containing information and statistics about
movies. IMDb was first published in 1990 by Col Needham, a computer programmer, as a group of scripts which
allowed users to search a list of film credits that were compiled with the help of a USENET group. The database
was incorporated in 1996. In 1998, Internet Movie Database Ltd. became a subsidiary of Amazon, which uses it as
a way to advertise movies and DVDs which it sells on its main website.
250 Artificial Intelligence Play (Ver 1.0)-X

