A simple method for unsupervised anomaly detection: An application to Web time series data

We propose a simple anomaly detection method that is applicable to unlabeled time series data and is sufficiently tractable, even for non-technical entities, by using the density ratio estimation based on the state space model. Our detection rule is based on the ratio of log-likelihoods estimated by the dynamic linear model, i.e. the ratio of log-likelihood in our model to that in an over-dispersed model that we will call the NULL model. Using the Yahoo S5 data set and the Numenta Anomaly Benchmark data set, publicly available and commonly used benchmark data sets, we find that our method achieves better or comparable performance compared to the existing methods. The result implies that it is essential in time series anomaly detection to incorporate the specific information on time series data into the model. In addition, we apply the proposed method to unlabeled Web time series data, specifically, daily page view and average session duration data on an electronic commerce site that deals in insurance goods to show the applicability of our method to unlabeled real-world data. We find that the increase in page view caused by e-mail newsletter deliveries is less likely to contribute to completing an insurance contract. The result also suggests the importance of the simultaneous monitoring of more than one time series.

Learning to Predict Page View on College Official Accounts With Quality-Aware Features

At present, most of departments in colleges have their own official accounts, which have become the primary channel for announcements and news. In the official accounts, the popularity of articles is influenced by many different factors, such as the content of articles, the aesthetics of the layout, and so on. This paper mainly studies how to learn a computational model for predicting page view on college official accounts with quality-aware features extracted from pictures. First, we built a new picture database by collecting 1,000 pictures from the official accounts of nine well-known universities in the city of Beijing. Then, we proposed a new model for predicting page view by using a selective ensemble technology to fuse three sets of quality-aware features that could represent how a picture looks. Experimental results show that the proposed model has achieved competitive performance against state-of-the-art relevant models on the task for inferring page view from pictures on college official accounts.

Empirical Investigation of Factors That Influence Website Performance

Rapid advancements in digital technology have had a significant influence on businesses' websites. Organizations with well-designed websites have the potential of attracting customers, generating revenues, and increasing market share. Nevertheless, many organizations that are investing billions of US dollars on websites and page development are not attracting customers or generating revenues, incomes, and profits as expected. The study focuses on bounce rate as a key performance measure of website effectiveness. The research focus is on the factors that influence bounce rate to provide insights to advertisers and websites' designers to predict the effectiveness and quality of online advertisements before these ads are shown to both web users and online visitors. The authors investigated the influence of page view, unique page view, average time on page, entrances, and percent exit on bounce rate. The study shows that unique pageviews, average time on page, entrance, and percent exit have a positive and significant effect on bounce rate and have practical and research implications.

Investors’ attention: does it impact the Nigerian stock market activities?

PurposeThe author investigates whether investors’ online information demand measured by Google search query and the changes in the numbers of Wikipedia page view can explain and predict stock return, trading volume and volatility dynamics of companies listed on the Nigerian Stock Exchange.Design/methodology/approachThe multiple regression model which encompasses both the univariate and multivariate regression framework was employed as the research methodology. As part of our pre-analysis, we test for multicollinearity and applied the Wu/Hausman specification test to detect whether endogeneity exist in the regression model.FindingsWe provide novel and robust evidence that Google searches neither explain the contemporaneous nor predict stock return, trading volume and volatility dynamics. Similarly, results also indicate that trading volume and volatility dynamics have no relationship with changes in the numbers of Wikipedia pages view related to stock activities.Originality/valueThis study opens new strand of empirical literature of “investors' attention” in the context of African stock markets as empirical evidence. No evidence from previous studies on investors' attention exist, whether in Google search query or Wikipedia page view, with respect to African stock markets, particularly the Nigerian stock market. This study seeks to bridge these knowledge gaps by examining these relations.

Pemanfaatan Social Traffic untuk Meningkatkan Page Views Pada Video “Gagasan Isengin Acil ! Leri Marah” Di Akun Youtube dan Instagram Dalang Pelo

In this digital era, new media such as Instagram and Youtube has become the most popular media, especially among teenagers. With the advanced of these new media, content creators have been using Instagram and Youtube as platform to express ideas, creativity and upload their works. One of favourable content video in Youtube is Dalang Pelo's animated content that is funny and interesting. The creator content that made the video animation, is Nur Alif Ramadhan. The author is interested in analyzing Dalang Pelo's animation because currently there are many video animations with similar content, but it still exist. The purpose of this study is to find out how publisher gets a high page view on Dalang Pelo's Youtube video. The theory used is the New Media Theory which contains the concepts of website traffic, social traffic, and high page views. The research method used is a case study with a qualitative approach. In this study, the author wants to find out how Dalang Pelo’s video utilizes social traffic strategy to increase page views on it’s Instagram and Youtube accounts through interviews, observations and documentation. The results of this study can be conclude that the creator of Dalang Pelo utilizes two social media which are currently widely used by the community. By converting these media strategically, they can get a high number of page views. Di era digital ini media baru seperti Instagram dan Youtube menjadi media yang paling banyak digemari remaja. Dengan kemunculan media baru ini membuat para konten kreator menggunakan Instagram dan Youtube sebagai tempat untuk menuangkan ide, kreatifitas serta mengunggah hasil karya mereka. Salah satu konten yang saat ini banyak diminati oleh masyarakat adalah konten animasi Dalang Pelo yang lucu dan menarik. Konten kreator yang membuat animasi tersebut adalah Nur Alif Ramadhan. Penulis tertarik untuk menganalisis video animasi Dalang Pelo karena saat ini banyak animasi yang serupa dengan kontennya, tetapi ia tetap mampu bersaing serta mempertahankan eksistensinya. Tujuan dari penelitian ini adalah untuk mengetahui bagaimana publisher mendapatkan pageview tinggi di video Youtube Dalang Pelo. Teori yang digunakan adalah teori Media Baru yang didalamnya berisi konsep traffic website, social traffic, dan page views tinggi. Metode penelitian yang digunakan adalah studi kasus dengan pendekatan kualitatif. Pada penelitian ini penulis ingin mengetahui cara Dalang Pelo memanfaatkan social traffic untuk meningkatkan page views di akun Instagram dan Youtube Dalang Pelo melalui wawancara, pengamatan dan dokumentasi. Hasil penelitian ini dapat disimpulkan bahwa kreator Dalang Pelo memanfaat dua media sosial yang saat ini banyak digunakan masyarakat dengan cara mengkonvergensikan media tersebut sehingga bisa mendapatkan jumlah page view yang tinggi.

Are Wikipedia statistics useful in evaluating public interest in neurosurgical diseases? (Preprint)

BACKGROUND Wikipedia, a free, semi-editable online encyclopedia, is currently the fifth most popular website worldwide. It is a leading source of medical information for the public. OBJECTIVE To evaluate to what extent Wikipedia page view statistics can assess the public interest in neurosurgical diseases. METHODS The Wikipedia Massviews statistics tool was used to find the top ten and bottom ten pages on Wikipedia under the category “Neurosurgery” from January 1, 2016 to December 31, 2018. The top ten pages were analyzed for page view correlations with time, the languages available, the top five redirects and the respective PubMed data. RESULTS The most popular neurosurgical pages on Wikipedia were: Subarachnoid hemorrhage, Neurosurgery, Idiopathic intracranial hypertension, Intracranial aneurysm and Laminectomy. The bottom pages were: International Subarachnoid Aneurysm Trial, Dandy's point, Stereotaxic atlas, Traumatic pneumorrhachis and Decerebellate. More popular pages were available in more languages with the most popular being English. Users more often accessed Wikipedia on their phone browser than on desktop; this trend was also seen over Wikipedia English overall. PubMed publications did not correlate to the page views with time. CONCLUSIONS Despite the impact of media clamor, Wikipedia statistics offer valuable insight into public health interests and show how users access this information. Our study demonstrates that the most popular neurosurgical topics on Wikipedia are for aneurysms, stroke, hydrocephalus and spine surgery. These topics align themselves with the most common neurosurgery issues. We encourage physicians clarify any questions for the common well-read patient. This allows for a more efficient physician-patient interaction and can help highlight subjects of confusion for future patients especially since the readability of articles on neurosurgery is low.

DAG-CPM Scheduler for Parallel Execution of Critical Jobs

MapReduce applications having multiple jobs may be dependent on each other such as iterative Page View application [2] performs the required operation in several iterations before generating the result. Each iteration is considered as single job. Conventional Hadoop MapReduce schedules the jobs sequentially, but not customized to handle multi job application. Also, it will not perform the parallel execution of the dependent jobs. This prolongs the execution time to complete all the jobs. Therefore a new scheduler DAG–CPM Scheduler uses the critical path job scheduling model, to identify the jobs present in the critical path. Critical path job scheduling is optimized to offer support for multi job applications, critical path job is a series of jobs, if execution of a job is delayed, then time required to execute all jobs will be prolonged. DAG–CPM Scheduler schedules multiple jobs by dynamically constructing the job dependency in DAG for the currently running job based on the input and output of a job. DAG represents the dependency among the jobs, this dependency graph is used to insert a pipeline between the output of one job as input for map tasks of another job and it executes the dependent jobs in parallel which results into a substantial reduction in the execution time of an application. Experimental analysis on the proposed approach has been carried out on Page View application on Academic and research web server log file, such as, NASA and rnsit.ac.in of 10 GB data set. PigMix2 is executed on 8GB data set. Experimental results reveal that the average execution time is decreased by 41% compared to Hadoop in respect Page View application and Execution speed is 37.7% faster compared to Pig and DAG–CPM Scheduler can run 24.3% faster when compared to DAG–CPM Scheduler without pipeline.