Agricultural Research Management Information System

Home
Research Summary
Search
About Us
- ARMIS
- Brochure
Contact Us
Report
User Request
Data Input
Help
- Operation Manual
  - PDF
  - Video
- Program Area & Commodity
We have reached 37600 number of research entries at this moment.

- Logout

Research Detail

Home
Research
Detail

Sentiment Mining Based Social Media Marketing Using Artificial Intelligence

M. A. Masud
Assistant Professor, Department of Computer and Information Technology, Patuakhali Science and Technology Uni versity, Patuakhali, Bangladesh

M. F. Rabby
Faculty of Computer Science and Engineering. Patuakhali Science and Technology University

G. M. S. Islam
Faculty of Computer Science and Engineering. Patuakhali Science and Technology University

M. M. Billah
Faculty of Computer Science and Engineering. Patuakhali Science and Technology University

Md Naimur Rahman
Lecturer, Dept. of Electrical and Electronics Engineering. PatuakhaliScience and Technology University

Abstract/ Executive Summary:

The evaluation of social media has made great opportunities for marketers. Marketers use Facebook pages, blogs, web advertisements and have shown distinct strategies. Consequently people are expressing their opinions about their product all over the cyber world. Sentiment analysis, also known as opinion mining, analyzes people's opinions and retrieves the hidden information behind a large volume of opinion which is useful for different purposes of marketing. Technically, social networks are not aware at giving the facility to analyze sentiments. Using artificial intelligence, a step by step approach of sentiment analysis is discussed here. This paper proposes a methodology for feedback analysis by two ways. The first one is rating analysis and second one is textual analysis. The rating analysis takes numeric opinion and calculates them. Textual analysis finds out the polarity of a textual opinion. Finally the methods are implemented by a mobile based social networking app 'Pronet', which performs these methods to ensure appropriate feedback from consumer to their companies as well as to advertise products to reach the doors of customers.

Keywords: Artificial intelligence, Sentiment analysis, Social network, Social media, Marketing.

Location: Department of Computer and Information Technology, Patuakhali Science and Technology University, Patuakhali

Start Date: 00-09-2012

End Date: 00-03-2013

Program Area: Knowledge Management

Non-commodity: Database Management

Objectives:

1. To technical modeling of opinion analysis in social media through the help of artificial intelligence (AI).

Methodology:

To construct a model for a system, using the sentiment analysis strategies into a social network can be done in two step: i. Average rating. ii. Textual analysis. Textual Analysis: Proposed Algorithm for Step by Step Approach. The process of sentiment analysis has step by step activities. Each step prepares result for the next step. l ) Collecting textual data or information from targeted sites or media. 2) Cleaning the extracted data before the analysis is performed. Identifying and eliminating non textual and unnecessary content from the textual dataset are performed here. 3) Appraising and extracting reviews and opinions from the textual dataset through the use of computational tasks. 4) Classifying the opinions as good, bad, neutral etc. 5) Representing the detected polarities after passing into required equations considering the need of investigators. A word can be objective or subjective; if it is subjective, it can carry a positive or negative polarity, with certain intensity. For example, "fantastic" is a strong positive word, and "questionable" is a moderate negative word. And certainly there are a lot of words that are neutral in terms of polarity, e.g., "white." Formally, given a word w, it is associated with two scores, a positivity score pscore and a negativity score nscore, such that 0 pscore, nscore 1 and 0 pscore + nscore 1. If both scores are 0, the word is a neutral one. An excellent lexicon, SentiWordNet , encodes such linguistic knowledge. A data mapping process is illustrated in Table 1 through a simple example sentence. Notice that in the PSV, each word is accompanied by its pscore or nscore, depending on its polarity.

Source of Information: J Patukahli Sci and Tech. Univ. 2013, 4 (2) : 93-104; ISSN 1996-4501

Article Link (Online Journal):

Funding Source:

Total Budget (Tk.) :

Achievement/Findings:

This paper proposes a methodology for feedback analysis by two ways. The first one is rating analysis and second one is textual analysis. The rating analysis takes numeric opinion and calculates them. Textual analysis finds out the polarity of a textual opinion. Finally the methods are implemented by a mobile based social networking app 'Pronet', which performs these methods to ensure appropriate feedback from consumer to their companies as well as to advertise products to reach the doors of customers.

Knowledge Management: Journal