Detection and Prediction of Users Attitude Based on Real-Time and Batch Sentiment Analysis of Facebook Comments

Abstract

The most of the people have their account on social networks (e.g. Facebook, Vkontakte) where they express their attitude to different situations and events. Facebook provides only the positive mark as a like button and share. However, it is important to know the position of a certain user on posts even though the opinion is negative. Positive, negative and neutral attitude can be extracted from the comments of users. Overall information about positive, negative and neutral opinion can bring understanding how people react in a position. Moreover, it is important to know how attitude is changing during the time period. The contribution of the paper is a new method based on sentiment text analysis for detection and prediction negative and positive patterns for Facebook comments which combines (i) real-time sentiment text analysis for pattern discovery and (ii) batch data processing for creating opinion forecasting algorithm. To perform forecast we propose two-steps algorithm where: (i) patterns are clustered using unsupervised clustering techniques and (ii) trend prediction is performed based on finding the nearest pattern from the certain cluster. Case studies show the efficiency and accuracy (Avg. MAE = 0.008) of the proposed method and its practical applicability. Also, we discovered three types of users attitude patterns and described them.

A typology of Jordanian consumers after Covid‐19: The rational, the suspicious, and the cautious consumer

“…In addition, these consumers were similar to the laggards in Rogers's theory because they were very skeptical of this new situation as well as of the news and media and were very slow in changing their behavioral patterns or vaccinating (Rogers, 2003 ). The cautious consumers' attitudinal views are similar to those of neutral‐attitude consumers (Ajzen et al, 1982 ), who are characterized by a neutral mindset and tend to wait for others to take the initiative (Brendl et al, 2005 ; Tran & Shcherbakov, 2016 ). Without a distinctive view on the pandemic, these consumers were similar to Rogers's ( 2003 ) early/late majority classification as they were willing to take the vaccine once it had been used by several people and proven safe.…”

“…Overall, the rational, suspicious, and cautious consumers all had different emotional outputs. Unsurprisingly, people with positive attitudes normally have positive emotional outcomes and are often confident, optimistic, willing to adapt, have a greater sense of responsibility, and are highly reliable and trusting (Brendl et al, 2005 ; Reeves, 2006 ; Tran & Shcherbakov, 2016 ). In addition, people with negative attitudes are often found to be angry, pessimistic, frustrated, and highly skeptical and doubtful (Brendl et al, 2005 ).…”

Sentiment Analysis using Improved Novel Convolutional Neural Network (SNCNN)

“…Tran and Shcherbakov [14] outlined an unsupervised clustering pattern on sentiment analysis in text identification and predication with positive or negative connection of Facebook comments. (i) Discovery of patterns with real-time sentiment text analysis.…”

Digital technologies for searching and processing unstructured Information in modern higher education

“…Then the model can be represented as a sparse matrix Sthe adjacency matrix of the graph G, consisting of the elements: (6) The matrix can simultaneously contain links of the form (i, j), and (j, i). Let Θsome set of topics, T(i, j, t) is the function of reference weight (j, i) in the theme t ϵ Θ (the so-called subject weight), deg(i, t) is defined as: (7) Expression (7) determines the sum of the weights of all outgoing links of the i-th vertex in the subject t, then the probability of a user clicking on the link (i, j) is determined by the following expression: (8) The matrix of thematic scales for links (i, j) can be written in the following form: (9) Thus, thematic ranking of the i-th vertex can be represented as a system of iterative linear algebraic equations: (10) where d is the attenuation coefficient, k is the iteration number, is the value of the rank of the j-th vertex during the k-th iteration k.…”

“…One of the features of the functioning of educational systems is the need to process large amounts of unstructured data, filter them and adequately interpret, which is a priority task for the modern system of higher education [7]. To fulfill the functions of processing unstructured data within the framework of the implementation of the concept of industrial Internet of things in a distributed automated system, the Web Mining technology is implemented, using the Data Mining methods to research and extract information from Web documents and services [8].…”