• Educational Technology International
• /
• v.15 no.2
• /
• pp.71-88
• /
• 2014

A variety of studies to predict students' performance have been conducted since educational data such as web-log files traced from Learning Management System (LMS) are increasingly used to analyze students' learning behaviors. However, it is still challenging to predict students' learning achievement in blended learning environment where online and offline learning are combined. In higher education, diverse cases of blended learning can be formed from simple use of LMS for administrative purposes to full usages of functions in LMS for online distance learning class. As a result, a generalized model to predict students' academic success does not fulfill diverse cases of blended learning. This study compares two blended learning classes with each prediction model. The first blended class which involves online discussion-based learning revealed a linear regression model, which explained 70% of the variance in total score through six variables including total log-in time, log-in frequencies, log-in regularities, visits on boards, visits on repositories, and the number of postings. However, the second case, a lecture-based class providing regular basis online lecture notes in Moodle show weaker results from the same linear regression model mainly due to non-linearity of variables. To investigate the non-linear relations between online activities and total score, RF (Random Forest) was utilized. The results indicate that there are different set of important variables for the two distinctive types of blended learning cases. Results suggest that the prediction models and data-mining technique should be based on the considerations of diverse pedagogical characteristics of blended learning classes.

• The Korean Journal of Applied Statistics
• /
• v.32 no.3
• /
• pp.391-404
• /
• 2019

Customers' fixed characteristics have often been used to predict customer behavior. It has recently become possible to track customer web logs as customer activities move from offline to online. It has become possible to collect large amounts of web log data; however, the researchers only focused on organizing the log data or describing the technical characteristics. In this study, we predict the decision-making time until each customer makes the first reservation, using Airbnb customer data provided by the Kaggle website. This data set includes basic customer information such as gender, age, and web logs. We use various methodologies to find the optimal model and compare prediction errors for cases with web log data and without it. We consider six models such as Lasso, SVM, Random Forest, and XGBoost to explore the effectiveness of the web log data. As a result, we choose Random Forest as our optimal model with a misclassification rate of about 20%. In addition, we confirm that using web log data in our study doubles the prediction accuracy in predicting customer behavior compared to not using it.

• Journal of the Korean BIBLIA Society for library and Information Science
• /
• v.23 no.2
• /
• pp.209-233
• /
• 2012

In this study in order to verify the overall information seeking behavior of the Web-based OPAC users, it was analyzed transaction log file for 7 years. Regarding Web-based OPAC information seeking behavior, it was studied from the perspective of information seeking strategy and information seeking failure. In search strategy, it was analyzed search type, search options, Boolean operator, length of search text, number of uses of word, number of use Web-based OPAC, number of use by time, by week day. Also, in search failure, search failure ratio, search failure ratio by search options, search failure ratio by Boolean operator were analyzed. The result of this study is expected to be utilized for OPAC system and service improvement in the future.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.27 no.6
• /
• pp.1431-1439
• /
• 2017

In the security field, log analysis is important to detect malware or abnormal behavior. Recently, image visualization techniques for malware dectection becomes to a major part of security. These techniques can also be used in online games. Users can leave a game when they felt bad experience from game bot, automatic hunting programs, malicious code, etc. This churning can damage online game's profit and longevity of service if game operators cannot detect this kind of events in time. In this paper, we propose a new technique of PNG image conversion based churn prediction to improve the efficiency of data analysis for the first. By using this log compression technique, we can reduce the size of log files by 52,849 times smaller and increase the analysis speed without features analysis. Second, we apply data mining technique to predict user's churn with a real dataset from Blade & Soul developed by NCSoft. As a result, we can identify potential churners with a high accuracy of 97%.

• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.6
• /
• pp.237-244
• /
• 2022

In this paper, we present VAE-based recommendation using online behavior log and purchase history to overcome data sparsity and cold start. To generate a variable for customers' purchase history, embedding and dimensionality reduction are applied to the customers' purchase history. Also, Variational Autoencoders are applied to online behavior and purchase history. A total number of 12 variables are used, and nDCG is chosen for performance evaluation. Our experimental results showed that the proposed VAE-based recommendation outperforms SVD-based recommendation. Also, the generated purchase history variable improves the recommendation performance.

• Journal of Information Technology Applications and Management
• /
• v.13 no.1
• /
• pp.1-16
• /
• 2006

This paper empirically analyzed consumer price search behavior using Web log data of a Korean web site for price comparison. Consumer click-stream data of the site was used to test the effects of price level, product category, third party certification, reputation of retailers on click behavior. According to the descriptive statistics, 67.4% of shopbot users clicked the offer which was the lowest price returned in a search. We found that third party certification and reputation of retailers were significant determinants of clicking the lowest priced offer from legit analysis. We also applied Tobit regression analysis to estimate the price premium of the two determinants, but only reputation of retailers was found to have price premium of 4.9%.

• The Journal of Korean Association of Computer Education
• /
• v.23 no.2
• /
• pp.53-64
• /
• 2020

As online learning forms centered on video lectures become more common and constantly increasing, the video-based learning environment applying various educational methods is also changing and developing to enhance learning effectiveness. Learner's log data has emerged for measuring the effectiveness of education in the online learning environment, and various analysis methods of log data are important for learner's customized learning prescriptions. To this end, the study analyzed learner behavior data and predictions of achievement by machine learning in video-based learning environments. As a result, interactive behaviors such as video navigation and comment writing, and learner-led learning behaviors predicted achievement in common in each model. Based on the results, the study provided implications for the design of the video learning environment.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.22 no.2
• /
• pp.225-238
• /
• 2012

Among the various security threats in online games, the use of game bots is the most serious problem. In this paper, we propose a framework for user behavior analysis for bot detection in online games. Specifically, we focus on party play that reflects the social activities of gamers: In a Massively Multi-user Online Role Playing Game (MMORPG), party play log includes a distinguished information that can classify game users under normal-user and abnormal-user. That is because the bot users' main activities target on the acquisition of cyber assets. Through a statistical analysis of user behaviors in game activity logs, we establish the threshold levels of the activities that allow us to identify game bots. Also, we build a knowledge base of detection rules based on this statistical analysis. We apply these rule reasoner to the sixth most popular online game in the world. As a result, we can detect game bot users with a high accuracy rate of 95.92%.

• The Journal of Bigdata
• /
• v.1 no.2
• /
• pp.1-8
• /
• 2016

User activity log includes lots of hidden information, however it is not structured and too massive to process data, so there are lots of parts uncovered yet. Especially, it includes time series data. We can reveal lots of parts using it. But we cannot use log data directly to analyze users' behaviors. In order to analyze user activity model, it needs transformation process through extra framework. Due to these things, we need to figure out user activity model analysis framework first and access to data. In this paper, we suggest a novel framework model in order to analyze user activity model effectively. This model includes MapReduce process for analyzing massive data quickly in the distributed environment and data architecture design for analyzing user activity model. Also we explained data model in detail based on real online service log design. Through this process, we describe which analysis model is fit for specific data model. It raises understanding of processing massive log and designing analysis model.

• Journal of the Korean Society for information Management
• /
• v.20 no.2
• /
• pp.27-41
• /
• 2003

Query logs are online records that capture user interactions with information retrieval systems and all the search processes. Query log analysis offers ad advantage of providing reasonable and unobtrusive means of collecting search information from a large number of users. In this paper, query logs of NAVER, a major Korean Internet search service, were analyzed to investigate the information seeking behabior of NAVER users. The query logs were collected over one week from various collecions such as comprehensive search, directory search and web ducument searc. It is expected that this study could contribute to the development and implementation of more effective web search systems and services.