<i>R</i> Package <b>clickstream</b>: Analyzing Clickstream Data with Markov Chains

Scholz, Michael

doi:10.18637/jss.v074.i04

Cited by 27 publications

(25 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Various packages mention MCs -related models in the CRAN repository, whereby a few of them will be now reviewed. The clickstream package (Scholz, 2016), on CRAN since 2014, aims to model websites click stream using higher order Markov Chains. It provides a MarkovChain S4 class that is similar to the markovchain class.…”

Section: Analysis Of Existing Dtmc-related Softwarementioning

confidence: 99%

Discrete Time Markov Chains with R

Spedicato¹

2017

The R Journal

125

View full text Add to dashboard Cite

show abstract

Section: Analysis Of Existing Dtmc-related Softwarementioning

confidence: 99%

Discrete Time Markov Chains with R

Spedicato¹

2017

The R Journal

125

View full text Add to dashboard Cite

show abstract

“…First, it has been used in previous studies in similar problem domains, e.g., Yasami et al [29], Muniyandi et al [21] and Mori et al [20]. Secondly, k-means is well suited for the problem at hand [5] and has commonly available implementations [23,27]. It should, however, be noted that the anomaly detection approach presented in the present study is technologically independent from the k-means clustering algorithm.…”

Section: Initial Labelingmentioning

confidence: 99%

Anomaly detection of event sequences using multiple temporal resolutions and Markov chains

et al. 2019

View full text Add to dashboard Cite

Streaming data services, such as video-on-demand, are getting increasingly more popular, and they are expected to account for more than 80% of all Internet traffic in 2020. In this context, it is important for streaming service providers to detect deviations in service requests due to issues or changing end-user behaviors in order to ensure that end-users experience high quality in the provided service. Therefore, in this study we investigate to what extent sequence-based Markov models can be used for anomaly detection by means of the endusers' control sequences in the video streams, i.e., event sequences such as play, pause, resume and stop. This anomaly detection approach is further investigated over three different temporal resolutions in the data, more specifically: 1 h, 1 day and 3 days. The proposed anomaly detection approach supports anomaly detection in ongoing streaming sessions as it recalculates the probability for a specific session to be anomalous for each new streaming control event that is received. Two experiments are used for measuring the potential of the approach, which gives promising results in terms of precision, recall, F 1-score and Jaccard index when compared to k-means clustering of the sessions.

show abstract

“…A method that fits well with describing users' behavior on the macroscale is analyzing users' behavior as a sequence of events [17]. This way of analyzing provides insight into the order in which the user navigates the system, whereas the more classical approach to log file analysis for health care information systems only describes quantities of usage (eg, number of times pages are visited, mean duration of visiting a page).…”

Section: Log File Analysismentioning

confidence: 99%

“…The probabilities are calculated by means of Markov Chain modeling, meaning that, in contrast to summarizing sequential patterns, the purpose is to predict future usage behavior. Probabilities can be calculated by using zero-, first-, and higher-order Markov Chains [17]. These orders differ in that the next page category is predicted only on the basis of the current page category that is visited by the user (first order) or on a combination of the current page category with the page categories that the user was visiting before the current page category (higher order).…”

Section: Heat Mappingmentioning

confidence: 99%

Exploring How Professionals Within Agile Health Care Informatics Perceive Visualizations of Log File Analyses: Observational Study Followed by a Focus Group Interview (Preprint)

Klooster¹,

Noordzij²,

Kelders³

2019

Preprint

View full text Add to dashboard Cite

BACKGROUND An increasing number of software companies work according to the agile software development method, which is difficult to integrate with user-centered design (UCD) practices. Log file analysis may provide opportunities for integrating UCD practices in the agile process. However, research within health care information technology mostly has a theoretical approach and is often focused on the researcher’s interpretation of log file analyses. OBJECTIVE We aimed to propose a systematic approach to log file analysis in this study and present this to developers to explore how they react and interpret this approach in the context of a real-world health care information system, in an attempt to answer the following question: How may log file analyses contribute to increasing the match between the health care system and its users, within the agile development method, according to agile team members? METHODS This study comprised 2 phases to answer the research question. In the first phase, log files were collected from a health care information system and subsequently analyzed (summarizing sequential patterns, heat mapping, and clustering). In the second phase, the results of these analyses are presented to agile professionals during a focus group interview. The interpretations of the agile professionals are analyzed by open axial coding. RESULTS Log file data of 17,924 user sessions and, in total, 176,678 activities were collected. We found that the Patient Timeline is mainly visited, with 23,707 (23,707/176,678; 13.42%) visits in total. The main unique user session occurred in 5.99% (1074/17,924) of all user sessions, and this comprised Insert Measurement Values for Patient and Patient Timeline, followed by the page Patient Settings and, finally, Patient Treatment Plan. In the heat map, we found that users often navigated to the pages Insert Measurement Values and Load Messages Collaborate. Finally, in the cluster analysis, we found 5 clusters, namely, the Information-seeking cluster, the Collaborative cluster, the Mixed cluster, the Administrative cluster, and the Patient-oriented cluster. We found that the interpretations of these results by agile professionals are related to stating hypotheses (n=34), comparing paths (n=31), benchmarking (n=22), and prioritizing (n=17). CONCLUSIONS We found that analyzing log files provides agile professionals valuable insights into users’ behavior. Therefore, we argue that log file analyses should be used within agile development to inform professionals about users’ behavior. In this way, further UCD research can be informed by these results, making the methods less labor intensive. Moreover, we argue that these translations to an approach for further UCD research will be carried out by UCD specialists, as they are able to infer which goals the user had when going through these paths when looking at the log data.

show abstract

R Package clickstream: Analyzing Clickstream Data with Markov Chains

Cited by 27 publications

References 11 publications

Discrete Time Markov Chains with R

Discrete Time Markov Chains with R

Anomaly detection of event sequences using multiple temporal resolutions and Markov chains

Exploring How Professionals Within Agile Health Care Informatics Perceive Visualizations of Log File Analyses: Observational Study Followed by a Focus Group Interview (Preprint)

Contact Info

Product

Resources

About