Abstract
This paper will serve as a proposal for a master’s thesis project in the field of artificial intelligence. By working to apply inductive reasoning to the concept of mining data in social media networks it will be possible to analyze user behavior in such an open forum through the use of family rules.
Keywords: proposal, thesis, data mining, social media, family rules
Inductive Learning as a Data Mining Technique for Social Media
Data mining is, quite literally, the process of mining for data in one or more databases for the purposes of extracting specific information (Hendricks, 2013). In this day and age of technological advances, one would be hard pressed to find a business outside of a mom and pop store that does not engage in some form of data mining. Working to understand the different benefits that data mining offers to the companies who opt to go this route is the first step in understanding the reasons why data mining is becoming essential in today’s society.
In recent years, the knowledge base regarding inductive learning has expanded by leaps and bounds; we know more information about this methodology than ever before (Aksoy, 2008). Inductive reasoning is the process of reasoning based on experience, perceptions, and observations in order to form a direct conclusion (Santarosa.edu, 2013). The inductive reasoning completed by humans is not that dissimilar from data mining completed by algorithms in order to obtain data from a database for analysis and using that data to form a specified conclusion.
Use your promo and get a custom paper on
"Inductive Learning as a Data Mining Technique for Social Media".
Rules family algorithms are used as a means of extracting a set of “rules for a collection of objects belonging to a given set of classes” (Aksoy, p. 52). If the element applies to just one class, it may be used to form a rule, but if the element pertains to more than one class, it is overlooked and the next element is analyzed (Aksoy, 2008). Once the algorithm goes through all of the data, assigning rules as the occasion warrants, it will determine if there is any data left without a class; if this is the case, the process will start again, working through until all elements have an associated class assigned to them (Aksoy, 2008).
Social media has been described as many things, and though the definition of what social media consists of continues to change and evolve with each passing day, it is still possible to obtain a working definition regarding that particular platform type (Cohen, 2011). Perhaps one of the best ways to define social media is to state that it is an interactive worldwide platform generated off of user based and user driven content serving to allow for communication at any time, from any location, with any location the world over; because of the fact that it is a worldwide platform, and due to the fact that content is user driven and user generated social media platforms have become something more – feeding grounds for marketing data (Cohen, 2011).
Through a review of the different data mining techniques available, it will be possible to see that through the combination of inductive reasoning, data mining, the rules family, and social media platforms, there is nothing that will hold corporations back; the hypothesis being tested is that this three pronged combination is the best method to utilize in order to obtain data from social media platforms as a means of information gathering.
The project will be completed in the form of a qualitative historical analysis, looking at past literature on the subjects in order to determine the most effective method to use, followed by a write up of those findings and a recommendation of how best to apply the concepts contained within the research to any social media platform, thereby working to provide the biggest possible hook of all for those businesses that would have an interest in fishing for customers on a social media site. Working definitions of all major concepts will be provided and an analysis of the methodology chosen will be provided. A literature review will be completed on the relevant data and a complete analysis of past data on the subject will, it is hoped, work to validate the hypothesis attempting to be proven. A comparison of the different data mining techniques will follow, along with the analysis of that comparison, wrapping up with a summary of findings and information for future research.
- Aksoy, M. (2008). A review of rules family of algorithms. Mathematical and Computational Applications, 13 (1), p. 51.
- Cohen, H. (2011). 30 Social Media Definitions – Heidi Cohen. [online] Retrieved from: http://heidicohen.com/social-media-definition/ [Accessed: 5 Nov 2013].
- Hendricks, D. (2013). How Businesses Can Benefit from Data Mining. [online] Retrieved from: http://www.tmcnet.com/topics/articles/2013/03/21/331429-how-businesses-benefit-from-data-mining.htm [Accessed: 5 Nov 2013].
- Santarosa.edu (2013). Introduction to Inductive Reasoning. [online] Retrieved from: http://online.santarosa.edu/presentation/page/?36971 [Accessed: 5 Nov 2013].
