Читать книгу Social Network Analysis - Группа авторов - Страница 16

This section describes the application of the SNA using the Python libraries to a real-world application. For instance, let us consider the sentiment analysis of the social users in the COVID pandemic scenario or predicting and tracing the contiguous diseases. With the enhanced development of technology, the expected data can be attained by just typing the required keyword in the search engine. The number of sites of social networking is capable of providing more informative data that assist in the evaluation of SN. The data needed for the analysis are gathered through the application of data mining concept in social network sites. The creators of the social media platform, like Facebook, Reddit, Twitter, afford the users with Application Programming Interface (API) that assist in gathering the expected data from the website. Application Programming Interface acts as a medium of communication between the server and the client. It helps the creators to extract the data available in one location to the other with the provision of a function that assist in copying the data. The working principle of API differs from one programming language to the other. The data gathering, preprocessing, classification are the important stages in SNA, and it is depicted in Figure 1.3. Data gathering is the first step to execute any work in data mining. The process of data gathering is a flexible task, and it relies on the particular subject of user interest. Initially, the raw data are accumulated from the social network by requesting the data with a precise keyword.

After gathering the data from the social network, the data are preprocessed to execute the processes, like prediction or analysis. Based on the application, the collected data are processed with the preprocessing stages, and the data can be categorized and visualized. Nowadays, in Python, the classifiers implemented for an application is mainly any kind of the machine learning classifier that acts as a supervised machine learning approach. The classifier requires proper training using the labeled training data, without which the performance of the classifier cannot be analyzed. One of the commonly used statistical classifier is the Naïve Bayes classifier, which is generally used to classify the sentiments of people in COVID pandemic conditions. Such kind of classifiers generally utilizes the publicly available data (from the communal media data) in an efficient way to perform a prediction or analysis or classification problems.

Figure 1.3 Flowchart of social network.