Human Centered Computing: Social Network Analysis, Behavior Analysis And Personalization
Abstract— Human centered computing has been used to analyze the user or human behavior towards computer and internet. By using internet, all the history or log data will be recorded. Therefore, all these behaviors can be analyzing to show the pattern of frequent search keywords and what can make users more attracted to use internet. So, this research can conclude that human centered computing has been used to determine user personality and behavior using either social network or search engine.
Keywords— Human centered computing, behaviour, social network, personality, internet
Introduction
Human-centered computing (HCC) also mainly focuses on the design, implementation, and deployment of computer system and computational technology to analyse and understands the interaction between human and information technologies. Besides, HCC also aims at changing computing with new method to design and develop the system based on user’s need with the understanding of human beings and related to human society and condition. HCC also has been applied into e-commerce to improve user’s experience while using internet. The research of the human-centered computing can be presented as a three-dimensional space that consist of computer, human and environment.
However, HCC is only not about the interaction, design process and the interface but it also introduces us more about new knowledge, technology (facilities) people and also any related things that ties them together. The emphasis of HCC is not only in the computer, but it is simultaneously in understanding on how the experience of use affects human behavior, understanding on how the technologies could affects social values and how they change the society. In this article, social network analysis and behavior analysis for personalization will be discuss.
Social Network Analysis
Social network rapidly growing in many areas include mathematics and computer sciences. The study of social network can be use in real life survey data for economics, sociology and women studies. What is social network analysis? A collection of techniques, tools and methods to map measure the relationship among people and organization. Therefore, HCC can leverage the social network in creating systems that help to overcome any problems that occur in social network. Theory represent nodes (actors, people or things) that connected by ties, edges or link (relationship or interactions). It commonly can be seen in social media and friendship networks. In this section, social network analysis will be discussing.
Business
In strengthening an organization, social network analysis (SNA) are very useful to map a formal organization charts which reflect the way on how an organization works, so that they can see the existing connections and help them to build a good team job. In addition, social media help an organization to understand their targeted audience. For example, an American gaming and media streaming website used social media analytics to define the part of their audience which is more interested in certain types of video games. At the same time, it also helps to build a good relationship between our business and our customers. Besides that, social networks are also playing an important role in an organization (Sineriz, n. d. ) as a medium in marketing product or services.
According to the figure below, more consumers are recommended to use the internet to find solutions. (Smith, 2017). An organization which is successful in their businesses, they know that using a social media as one of marketing methods through advertisement of their products or services makes total sense because using social media can lower ad costs compared to traditional advertising methods such as print brochures and TV or radio advertising. With digital advertising, it is not only dependable, but it is also cheaper. Furthermore, by using a traditional advertise, it is difficult for the organization to convey the advertisement of their product to the targeted audience compare to the way on how social media work in an advertising. Other than that, social media advertising is able to help us constantly keep on track of how far our advertisement is performed and we are able to manage our advertisement on the fly and instantaneously we can see the results as soon as possible.
Social Connections
Human center computing has connection with social network analysis. It is a method to support and enhance social interaction. We can identify some application that are close to human such as Facebook and twitter to more traditional social networks of people in a small town. Do you ever think how internet knows what product you want to buy or who to add as a friend? It is because social overlaying network that connect and transfer friendships, information, money and power, you start to see how analyzing things through the analysis of social network can lead to new realization about culture, politics, history and lots of other interesting topics. Social network analysis can generate graphical representation that reveal individuals in populations that bridge social groups.
For example, adding geographic mapping data to understand how physical environment chance network dynamics. For mathematicians, there still space for creating new equations or algorithms. The power of social network analysis can be seen in common media such as Facebook, twitter in figure below.
Behavior analysis for personalization
HCC in behavior analysis for personalization represent the frequency of the activity between end-user with computer. This behavior analysis can be identified the patterns of end-user performances in computer. There a lot of method to identify this behavior analysis and personalization which is in an online time over a data stream.
Patterns of Analysis
The patterns from the analysis could be used for frequent changes and increasing trend of sites. There many approaches for discover the pattern using graph, rules or model but mostly it identifies the patterns offline. The output of the patterns will be outdated and limited due to the short-term tasks. The proposed method for behavior patterns and personalization using online time between end-user and website. Design of pattern mining can be used in an online time. Each of activity, set of the action will process in single pass in continuous stream. Previous behavior is clustered and stored in models. These models will be updated continuously each session of activity and it can quickly react to the changes of data. To prevent the high memory consumption, the session of activity will store with limited capacity. This paper used stream clustering to identify the groups of similar activity continuously over time and pattern mining performed for each group. These pattern mining will generate recommendation-based user actions.
Data collection
HCC can be characterized as a multistage problem-solving process that requires designers not only to analyze and foresee how users are likely to use an interface but also to test the validity of their assumptions regarding user behavior in the real world. Data collected through HCC can be used for personalization. HCC application that has been emerged in recent years is the automatic analysis of human activities and their social behavior. Behavior analysis and personalization is a process to determine what the web looks like today debatably more than any other technology. Personalization means meeting user’s requirement efficiently, making interaction faster and easier and increase user satisfaction for probability repeat visit. Web personalization is an automated process to identify consumer, collect user’s behavioral record, analyzing user preferences and tailoring content for each user.
In other words, HCC for web personalization has been adopted in online shopping mobile application to understand user’s behavior and their frequent visit site. Besides, it is also important to improve user’s experience and consumer loyalty by giving some suggestion items based on their recent or frequent search item. There are two types of web personalization which are Before-search Web Personalization (BSWP) and After-search Web Personalization (ASWP) to provide better recommendation and to satisfy users’ different need. By using these technique, analysis of user access pattern can be done to identify user web-browsing behavior from online website. Using user’s behavior, it has been used to preprocessing the log-file that contain the pre-fetch, predict and the improvement in the performance on built a web server. All the data that has been collected during the feedback can be use and recorded in the future work for improve the website.
Tracking User Behavior
To know and do the analysis about user or human behavior, it is can be tracked user’s behavior offline or while the system is idle. This process can be tracked during the system that going offline or when their processor or the computer got the poor workload. Using these data, we can track their personalization on the activity that they did when they want to browse the website. However, there are some disadvantages to offline website where some time the data will be limited on which information they can published or updated. The data can be track while the website or application is on online mode. For example, all the information that the user wants to update on their social network or other website, they need to be online to publish all the information. This information’s that has been shared using online way, it can be exchanged or shared with other people. Therefore, by sharing and exchanged the information, the data can be tracked, and it can do the analysis of the behavior for personalization. This data also can be recorded using session identification. It is to discover the activity or time that has been consumed by the user while browsing certain web site. Each web site that has been visited by the user, it will be recorded as a personal and individual session. For example, when user visit airlines website, the time that has been spend by user will be counted. However, once the session request for additional time, it will be considered as a new session even though for the same user.
Pattern Discovery
More user behavior pattern can be discovering from the information that has been collected by applied mathematical analysis, serial patterns and classification. It is not only limited to the technique of online and offline website and web usage mining. By applying those methods, we can use variety of approach to find patterns in behavior of user for personalization.
Conclusion
With all the methods and explanation above, human centered computing can be applied in to both social network and web browsing to find and analyze user behavior pattern and personalization while using Internet.
