Facebook an Anti-Stereotyping Tool: A Case Study

Facebook, the most popular social media (SM) platform has penetrated every nook and corner of the world. SM is now treated as the ‘fifth Estate’, other than legislative, executive, judiciary, and mainstream media. The power of SM as a critique is widely acknowledged. Establishments are finding it difficult to deal with it at times. Due to its ease of usage and relative anonymity, the general public finds it very convenient to put across their viewpoints, even if it’s against the establishment. Some establishments at times are at loggerheads with champions of freedom of speech including civil rights activists. SM has been used for propaganda, marketing, and awareness campaigns. In this paper, we are proposing to use this powerful tool towards social change. Through a case study, a detailed process is being proposed for using social media particularly Facebook as an an-ti-stereotyping tool. The response to an online survey, the outcome of opinion mining, and the enthusiastic response to our case study by the targeted audience validate our hypothesis that Facebook can be effectively utilized as an anti-stereotyping tool.


Introduction
In 1954, the term "Social Network" was coined by Barnes [1]. With the advent of Web 2.0, social networking sites exploded and attracted unprecedented interest from people all over the world [2]. The community driven uprisings including Arab  [2]. Among the social media platforms Facebook is the most popular [3]. Impact of social media in consumer perception management is widely studied. The social media can be successfully used for propaganda and perception management [4]. Social media can become an effective alternative tool for social mobilization [5]. The Pangals have been in Manipur since 615 A D, although the majority of them arrived at 1606 AD and were given the local Meitei, the original Manipuri women to marry [6]. The Pangals have migrated to other states and also to Bangladesh during the seven years devastation (1819 -1826 CE). The Pangals follow Islam and they have a unique and original culture which has resulted from adopting the local Meitei culture so the Pangals become one of the original inhabitants of Manipur.
The changing status of Pangals over the period of time and the existence of the various circumstances of stereotyping of Pangals is the reason why the study of the community is particularly undertaken. During the king Khagemba's reign there were great agricultural developments and Manipur society was more diverse and plural with the settlement of Pangals. Pangals contributed to the enrichment of the economy and cultural life of Manipur. Thus, Pangals enjoyed an equal social status with other communities [6]. Of late, Pangals started facing discrimination, marginalization, stereotyping from political leaders and fellow Manipuri. The common stereotypes held by Meitei and other non Pangals is that they are anti socials and prone to thievery. There are several instances of stereotyping against this community. In 1993 Pangal Massacre,130 Pangals were killed [7]. In 2016 Yangbi Garden, Mayang Imphal incident, two Pangal students were killed. In March 2016 Holi incident, six Pangal boys were beaten up [8]. Md Hasmad Ali, a primary school teacher was killed alleging as cow thief [9]. In April 2018  [11].
In this paper, we present a case study of Facebook as an anti-stereotyping tool. The Pangal community of Manipur is taken as the subject of study. We have tried to change the general perception about the Pangals using Facebook thus validating Facebook as an anti-stereotyping tool. The remaining part of the paper is organized as follows. Section II describes the literature survey to identify uses of Facebook in different domains. Section III describes the methodology used in the case study in step by step procedure. Section IV introduces social media particularly the features available in Facebook. These features are utilized in the case study. It also covers the text classification aspects. This will help us in understanding the sentiment analysis of the Facebook comments. Section V discuss the response to the online survey and the outcome of the sentiment analysis. Section VI gives out the result of the case study. Finally, the paper is concluded by giving the future scope of work.

Literature Survey
The Literature survey follows the accepted method of Keyword Search from Peer Review Journals and Conferences reports.
The most cited ones whose analytical deductions are available are being sorted out. It also includes certain books and government reports. It is basically to bring out the various uses of Facebook in different domains and how is the impact of Facebook in a particular domain is studied. [4] has done a qualitative case study and examined how social media(SM) was used as a platform for propaganda warfare waged by clandestine bloggers and special operations commandos (known as trolls). It uses manipulation of public perceptions of events by controlling an element of rhetoric known as narratives. [5] examines how SM could be deployed for transformational mobilization and development. Role of SM in promoting social mobilization as a means of development in Nigeria was studied. [12] addresses the use of Facebook as a learning tool for higher education students, with emphasis on developing countries such as Malaysia. The purpose of their study was to identify students' per- [14] introduce a framework for SM impact, which included the most important role in enhancing the deployment of SM in crisis in order to minimize the negative impact on education's sustainability. This framework can be used as strategic decision-making tool for the analysis of the gaps and inefficiencies in any SM plan that is deployed and the management challenges arising from the pandemic. [15] presents a comprehensive review of recent studies that employed Facebook as a tool for teaching and learning in institutions of higher education. [16] present a SM enabled social movement the 'Electoral Reform' in Malaysia and significant role of SM in empowering citizens towards change in their community. [17] present findings from interviews with 26 small NPOs' SM professionals on how they use multiple SM sites to support public engagement.
Small NPOs use multiple SM sites to engage with different stakeholders towards various ends. [18] presents a case study of a voice-based community media platform in rural central India and the experiments it has undergone with multiple community mobilization strategies over a period of five years. They analyzed different phases of community mobilization and drew insights related to how technology platforms could be appropriated by special activists to drive their own agenda. [19] describes the module for Analysis of social network Facebook Graph API. The study identified the areas of interest of the user, analyzed comments in Facebook to identify trends. [20] carries out a study on social influence produced by the contents pub- Most of the paper reviewed examined the behavioral side of social media, investigated the aspect of reviews and recommendations. The theories used were user and gratification theory, social capital theory and goal setting theory. People have studied about using Facebook for mobilization, awareness campaign, as a communication tool, and as a marketing tool. Impact of social media on social moments have also been studied.

Research Methodology
The research methodology used is mixed of qualitative and quantitative approach. Data is collected through Facebook scrap-

Inspiring Stories Creation
The original inspiring stories of the pangal community were obtained from two sources. The primary and secondary sources.
The primary sources are the individual himself/ herself, family and friends and general public. The method of collection of data is through WhatsApp or e-mail. The secondary sources consisted of books, papers, news reports, and online resources.
These were directly collected and properly edited. It was ensured that every positive and inspiring aspects of the profile was adequately highlighted along with testimony. The testimonies were authentic records and original artefacts. All the existing records with respect to the profile were used to make it more appealing to the general public. The presentation of the profile in the Facebook page update was with a suitable catchy tagline. An example of the profile creation is given below as an illustration.

Study of the performance of the Facebook Page
Certain readymade statistics are available with the Facebook page business suite tools. These statistics are freely available, and it can be used to ascertain the general response to the updates of the page. It also gives insight into the category and geolocation of the followers. The engagement and reach of the page is also captured.  • Socializing

Collection of Comments for Opinion Mining
• Entertainment • Self-seeking • Information

Facebook Page
Page is the primary system of Facebook through which users are exposed to content posted by people across the globe and the news feed. Facebook selects around 1500 updates in the news feed. Facebook page is attached to a personal profile so Facebook knows who owns the page but pages are designed for business or for a cause. It has a number of features available including analytics, advertising and post scheduling.
Six of the options for creating a page : • Local business or place.
• Company organisation or Institution.
• Brand or Product • Artist, Band or Public figures.
• Causes of Communities.
Each page type has a dropdown menu with and extensive list of categories. The profiles of individuals are updated in the page at regular intervals.

Text Classification
Text classification is the categorization of the text in some predefined categories. A classifier takes the text input and analyze it then automatically assigns them in the predefined categories. Mainly three approaches of text classification are available in Natural Language Processing (NLP). • Machine Based Classification: Based on past observation on datasets and continuous learning.
• Hybrid Based System: Uses combination of both Rule based and Machine based methods.

Machine Based Classification
The steps involved in Machine Learning text classification are as follows: • Pre -processing: Unlike structure data, features are not explicitly available in text data. Thus, pre-processing of the raw data is required.
• Feature extraction: Feature extraction can be done in various ways. The common methods are bag of word and word embedding method.

• Prediction
Popular Machine Learning Algorithm used in Text Classification are Naïve Bayes group of algorithm, Deep Learning, and Support Vector Machine. For feature extraction common method is bag of word method.

Bag of Word Method
In bag of word method, a dictionary of all the word is created. It first converts each document into a vector. There are three ways to convert word into vectors as given below: • Count vector method: In this we count the occurrences of each word in the document. Thus, what we are doing is any vector that represents a text that has to contain information about the probabilities of the appearance of certain words within the texts of a given category, so that the algorithm can compute the likelihood of the text's belonging to the category.

Support Vector Machine (SVM)
It is another powerful text classification Machine Learning Algorithm. It requires more computational resources, but results are faster and more accurate. SVM draws a line/hyperplane that divides a space into two subspaces, one subspace contains vectors that belongs to the category, and another subspace contains vectors that do not belong to that category. The best hyperplane is the one with the largest distance between each category.

Deep Learning
These are sets of algorithms and techniques inspired by how human brain works. Two main deep learning architectures for text classification are Convolutional Neural Network (CNN) and Recurrent Neural Network (RNN). Deep Learning requires more training data.

Sentiment Analysis using VADER
Valance Aware Dictionary for sentiment Reasoning (VADER) is a lexicon and ruled based approach developed by C J Hutto and Eric Gilbert in 2014 [22]. Although it is domain agnostic, it is especially suitable for Social Media Analytics. The advantages of VADER are: • No training data required     Table 1].

Conclusion
A methodology to use Facebook as an anti-stereotyping tool have been proposed. In order to validate the methodology, a way of online survey and sentiment analysis is suggested. The response and the engagement of the Facebook page created for the purpose shows a very positive outcome. It is strongly observed that Facebook definitely can be used as a social change tool. The method to carry out analytical measurement of the degree of achievements of the desired change in the society can be developed in future. The parameters how the changes have to be measure needs to be defined to have a foolproof methodology. In future the process can be streamlined and proper analytical validation methodology in order to measure the degree of effectiveness of Facebook as a social change agent can be developed. Some suggested areas where it can be used are given below. However, the possibilities of utilizing the process is manifold.
• Perception Management