Inspecting Student's Behavior When Using Online Services website by Means of Data Mining and Text Mining Techniques
Introduction
In the recent times, the web-based systems that related to education area are developed to record students’ online interaction manners. Also, those web-based systems mount up big volumes of data that offer researchers as well as decision makers a goldmine of details on student's behaviors, characteristics, and patterns (Abdous and He, 2011). Accordingly, various automatic and electronic analyzers have been developed to investigate the big amounts of data (Chiang et al., 2011).
The release of data mining applications and techniques has made it easier to analyze students’ interaction and behavior when using web-based systems (Romero et al., 2011). Data mining has become the most dependable technique that automatically extracts implicit and explicit models from great amounts of data (Klosgen and Zytkow, 2002).
In the educational field, data mining offers the educational organizations the competency to discover, envisage and study abundance of data collection to retrieve valued samples of students’ manners instead of resorting to old survey approaches (Abdous and He, 2011). Furthermore, the educational organizations can improve their operational and education practices based on the useful information retrieved from big data processing. Consequently, educational data mining has become an essential research field with a particular concentration on exploiting the ample data produced by different organizations that related to education erea towards improving teaching, learning and decision-making making (Liao, et al., 2012).
Educational Data Mining (EDM)
A number of leading Educational Data Mining professionals (e.g. Romero and Ventura, 2010) categorize the tasks of EDM into a few classifications such as visualization and statistics, prediction, clustering, mining of relationship, detections of the outlier, and mining of text. EDM can be utilized to evaluate students’ educational behavior, develop the educational process as well as monitor students’ learning. In addition, it is used to deliver feedback and adjust educational recommendations depending on students’ educational activities, to assess educational resources. Finally, it provides the decision makers of educational organization with good information that help them to spot any abnormal educational interactions and complications and to accomplish more profound realization of educational progress.
Methodology
We are targeting an educational organization at the Saudi Arabia where we collect the major data necessary for the study such as the students number, the subjects number taught at the university and the types of services provided by the educational website. We are investigating the types of activities the students make on the website beside the way they interact with their teachers and the management. Moreover, we are investigating the type of text messages the students write, and the replies sent back to them.
The researcher wrote a PHP code to parse the retrieved texts in order to get the questions and suggestions of students into a database of Microsoft SQL Server and successively transferred the data to excel spreadsheets to facilitate data analysis. Data was pre-processed to remove the test messages of the system that posted by staff members.
For this study, we adopt the model of Abdous and He, 2011.
our study contains 2 parts. Throughout phase 1, data was pre-processed, in which it was converted into an operational layout, primarily by cleaning, allocating attributes, and assimilating data. Afterward, various data mining procedures were applied as well as text mining methods to inspect the two diverse data groups to achieve insights regarding students’ usage of the website and educational behaviors.
The inquiry searches are principally utilized to investigate the ideas of students, uncover their stimulating manners, networking (with teachers/management) and infrequent information. The objective of clustering analysis in this part is to spontaneously categorize the posts of student according to the essential content when using the website. The analysis unit during clustering analysis is based on the places the student visit in a website and the feedback they provide for their teacher/management. Garrison et al., (2010) stated that there are challenges when selecting a suitable analysis unit for the texts.
Many researchers employed grammatical units like sentences and passages as well as operational units like complete posts and threads for the analysis unit (Zha and Ottendorfer, 2011). Different ways have different strengths and weaknesses (e.g., reliability issues).
our study utilizes a message-based mining method to analyze texts automatically without the traditional exhausting methods. Accordingly, the line level unit is the best fit to accomplish the purpose of the study. Study of Garrison et al. (2001) has revealed that the message-based mining method is helpful because it gives more details on the behavior of its writer and thus helps to develop the systems accordingly.
Findings
The finding of study has shown inconsistencies as well as resemblances in students’ usage of the online services and giving written feedback. The students were supposed to be more active in contacting their teacher; however, they intended to contact the management most of the times.
The study shows that the analysis contributes to the decision-making process in the educational organization; in terms of development and problems resolution. The algorithm developed to represent an ongoing source of information for the educational organization. The updates made upon the website do not affect the algorithm functions.
Conclusion
The current outlined model can be applied to another website in order to translate the big data generated by them into meaningful information that could be used by the decision makers.
References and work Cited.
Abdous, M., & He, W. (2011). Using text mining to uncover students' technology-related problems in live video streaming. British Journal of Educational Technology, 42(1), 40’49.
Chiang, D., Lin, C., & Chen, M. (2011). The adaptive approach for storage assignment by mining data of warehouse management system for distribution centers. Enterprise Information Systems, 5(2), 219’234.
Garrison, D., Anderson, R. T., & Archer, W. (2010). The ‘rst decade of the community of inquiry framework: A retrospective. The Internet and Higher Education, 13, 5’9.
Klosgen, W., & Zytkow, J. (2002). Handbook of data mining and knowledge discovery. New York: Oxford University Press.
Liao, S., Chu, P., & Hsiao, P. Y. (2012). Data mining techniques and applications ‘ A decade review from 2000 to 2011. Expert Systems with Applications, 39, 11303’11311.
Romero, C., & Ventura, S. (2010). Educational data mining: A review of the state of the art. IEEE Transaction on Systems, Man, and Cybernetics, Part C: Applications and Reviews, 40(6), 601’618.
Zha, S., & Ottendorfer, C. (2011). Effects of peer-led online asynchronous discussion on undergraduate students’ cognitive achievement. American Journal of Distance Education, 25(4), 238’253.