Search

Real-Time Bidding (RTB) is an important mechanism in online search advertising involved with the interaction among platform, advertisers, and customers, where the proper bid for each page views plays a crucial role for good marketing results. Enhanced Cost Per Click (ECPC) bidding is one of typical bidding strategies in the RTB mechanism where the advertisers will raise bid prices for traffic with high conversion rates under the premise of total advertisers' cost control in different time periods of the day. The bid ratio in ECPC as a correction coefficient is used to control the advertising cost of advertisers. However, the optimal bid ratio is hard to be derived due to the complexity and volatility of the auction environment. To address the challenges, we formulate the hourly-aggregated level of ECPC bidding as a Markov Decision Process with innovative state and reward function according to business. Besides, quite different from prior work that has to construct an auction model-based environment to solve the MDP problem, we propose a novel framework based on model-free reinforcement learning that does not rely on the agent's explorations in the environment. We furthermore design an adaptive action choosing strategy which adjusts the bidding behavior dynamically and improves the performance of categories/groups's consumption qualified rate. Experimental results on real dataset demonstrate the effectiveness of our framework from the perspective of consistency, stability and online effect.

In this article, we consider the problem of assessing the cross-layer performance of wireless centralized communication networks (e.g., 5G and LTE). We consistently introduce a user web browsing model and find the relationship between the numerical characteristics of a wireless network and the network's perceived quality when users browse the pages. An optimization problem was formulated based on the results obtained. Solution of optimization problem determines the lower bound of the average waiting time for loading additional page objects for given page view probabilities and the average waiting time for the main page object loading using various scheduling algorithms at the data link layer.

Electronic journal (e-journal) management comprises several aspects, specifically pageviews, sessions, visitors, and new visitors. Sessions or the number of unique visitors from a journal page is an essential indicator of a journal's outcome. Therefore, it is necessary to forecast the number of unique visitors to determine the strategy for developing a journal. Thus, it is expected to be able to accelerate the journal accreditation system in the future. In this study, this paper predicts the number of unique visitors to the journal by developing a time series forecasting model. Forecasting was done by applying the Backpropagation. The method has the advantage of being able to adapt to changes that occur in the input and output values. There are three time series data input models for this research, specifically three days, seven days and 14 days. The accuracy of forecasting results was measured using the MAPE evaluation of several forecasting models and BPNN architecture. The results show that the best forecasting is using forecasting model 1 and architecture 2-5-1 with an accuracy value of 69.9%. Thus, the performance of the Neural Network in this study is relatively good.

The e-commerce era is witnessing rising new arrivals of items on e-commerce platforms every day. Identifying potential popular items accurately is of great importance in creating commercial value. Click-Through Rate (CTR) is a general indicator to evaluate item popularity. However, existing methods fail in new arrivals prediction because of sparse item features, missing item statistics and high time complexity of computing for all pairs of users and items. To tackle these challenges, we propose a novel Adversarial Two-tower Neural Network (ATNN) model for new arrivals CTR predictions by introducing an adversarial network to a two-tower network. We design a generator and a discriminator to better learn an item vector based on item profiles without item statistics. We also develop a strategy with an $\mathcal{O}(1)$ time complexity for a new item’s popularity prediction by constructing a user group and utilizing its mean user vector in a time-efficient manner. We implement ATNN on a large-scale real-world dataset from one of the world’s largest e-commerce platforms, "Tmall.com". Empirical results show that ATNN is strongly capable of learning item vectors from item profiles for e-commerce. Furthermore, by introducing multi-task learning technology, we extend ATNN to food delivery service. Experimental results on one popular food delivery platform, "Ele.me", demonstrate that ATNN can recognize attractive and welcoming new restaurants that have higher Value per Page View (VpPV) and generate more Gross Merchandise Volume (GMV).

In this work, we investigate the resource allocation problem in contract display advertising with a learning and planning approach. We formulate the problem as a mixed integer nonlinear programming, aiming to determine the resource allocation strategy in each advertising slot with CPC (cost per click), CPA (cost per action), CPM (cost per mille), CTR (click through rate), and budget constraints, thus maximizing the total page view. Since that the page view (PV) prediction is extremely complex, we assume our PV prediction function is concave and then we learn the PV prediction function for each slot using historical data. The concave assumption makes our objective function nonlinear, and we employ a piecewise linear approach to approximate PV function. To test the proposed model, we conduct numerical experiments with real data from the company. Results reveal that the proposed model outperforms the current allocation strategy. Moreover, our approach significantly increases the total contract page view and lowers the overall CPM.

In this paper, we propose a method that implicitly contains multiple features to describe the relationship between two variables, which was referred to as trend prediction. This method takes into account the relationship between the slot page view and multiple features. The data of different advertising slots are constructed by predicting methods such as tree model and neural network with category feature densification by auto-encoder model. To reduce the smooth time in the prediction curve, we take the logarithm function as a priori to regress the predicted curve. Finally, compared with our previous prediction model, the trend prediction model effectively improves the overall page view under the unconstrained mixed-integer programming model.

Extracting useful and relevant information from the massive amount of data by web users becomes a challenging task. This work regarding applications based on the use of Web Usage Mining (WUM). Clickstream, transaction data, and user profile data represent various sources of web usage data. Preprocessing is an essential process in WUM for understanding the user's life as a whole within a session on a website. It signifies the aggregation of consequent series of page views executed by a singular user traversing through a website. Therefore, it represents a critical task that involves converting raw web log data to obtain a suitable pattern for efficient analysis. The importance of these processes relies on the complex nature of web architecture, and it takes almost 80% of the mining process. This work introduces a complete framework for data preprocessing with a comprehensive demonstration of a real dataset.

In support of the mission to ‘increase interest, enjoyment, confidence, and ability in mathematics and computer science among learners and educators in Canada and internationally’, the University of Waterloo’s Centre for Education in Mathematics and Computing (CEMC), working in collaboration with local education technology provider DigitalEd, made its mathematics content globally accessible to serve as an online complement to mathematics educators, in addition to providing a source of online resources for independent study and upskilling of non-registered learnersIn September 2019, the full library of resources was launched and promoted globally, grouped into courses based on the topics covered in the Canadian curriculum for grades 7 to 12, along with special topics in financial literacy. As of April 2020, the resources had been accessed by over 250,000 users across 213 countries, with over 1.7M exercises attempted. Peak usage saw over 300,000 pageviews in a day, reaching over 40,000 unique users across 86 countries, with over 99.9% uptime, utilizing dynamically scaled cloud instances to handle unexpected loads. Continuous system monitoring was employed as a supplement to automated alerts and reliability engineering tools in order to maintain system performance and uptimeThis paper provides an overview of the considerations and principles followed in the planning and design of mathematics learning resources, structured as year-long courses in secondary school mathematics, and provides a summary of the requirements for the infrastructure put in place to enable continuous, scalable, global open access. It concludes with an overview of recommendations for localization as a means to increase adoption in regions where English is not the primary educational language.

Clickbait is a bad habit of today’s web publishers, which resort to such a technique in order to deceive web visitors and increase publishers’ page views and advertising revenue. Clickbait incidence is also an indicator for fake news and so, clickbait detection represents a mean in the fight against spreading false information. Recently, both the research community and the big actors on the WWW scene such as social networks and search engines, turn their attention towards this negative phenomenon that is more and more present in our everyday browsing experience. The detection techniques are usually based on intelligent classifiers, features selection being also of great importance. This paper aims to bring its own contributions in clickbait analysis and detection by presenting a new language independent strategy for clickbait detection that considers only general features that are non language specific. This approach is justified by the need for a higher level of abstractization in the clickbait detection, allowing its usability for articles written in different languages. A complex experiment on a real sample dataset was conducted and the obtained results are compared with the most relevant previous work results.

To create analytical models for disseminating information in social networks, a special technique has been implemented for selected statistics in the network VKontakte (VK). The statistical features of the change in the number of subscribers to social groups in terms of the pageviews are investigated. The classification of the groups of social network users is proposed; their differences and common patterns of mutual transitions are determined.

The tourism industry plays an important role in Thailand's GDP which has given a huge economic boost every year. As we know, most of the tour operators own websites that are accessible worldwide and usually measured by the performance of online marketing. Therefore, the purpose of this paper is to develop a methodology to analyze the effects of sessions on unique visitors and unique page views of tourism websites based on time series data on Google analytics. This cannot be defined by simple regression as it depends on multiple factors, i.e., visitor types, traffic(channels) and technology (browsers). Besides, this paper describes how to use the most accurate data to gain an effective result. The results can be applied to other tourism websites for their analysis as well.

Digital Repository Systems have been used in most modern digital library platforms. Even so, Digital Repository Systems often suffer from problems such as low discoverability, poor usability, and high drop-off visit rates. With these problems, the majority of the content in the digital library platforms may not be exposed to end users, while at the same time, users are desperately looking for something which may not be returned from the platforms. The recommendation systems for digital libraries were proposed to solve these problems. However, most recommendation systems have been implemented by directly adopting one specific type of recommender like Collaborative-Filtering (CF), Content-Based Filtering (CBF), Stereotyping, or hybrid recommenders. As such, they are either (1) not able to accommodate the variation of the user groups, (2) require too much labor, or (3) require intensive computational complexity. In this paper, we design and implement a new recommendation system framework for Digital Repository Systems, named dpSmart, which allows multiple recommenders to work collaboratively on the same platform. In the proposed system, a user-group based recommendation strategy is applied to accommodate the requirements from the different types of users. A user recognition model is built, which can avoid the intensive labor of the stereotyping recommender. We implement the system prototype as a sub-system of the FIU library site (http://dpanther.fiu.edu) and evaluate it on January 2019 and February 2019. During this time, the Page Views have increased from 8,502 to 10,916 and 10,942 to 12,314 respectively, compared to 2018, demonstrating the effectiveness of our proposed system.

In recent years, tourists have been relying more on information from online sources for planning and managing their trips. The local tourism communities engage these changing behaviours by using information technology and social media for better connection with customers. Tourism agencies use several online channels such as a website, a social media application, and manage their own contents. For Thai local operators, this approach could be quite costly and time consuming. In case of the Pongyeang tourism community, there is a demand for a community website and a social media page for stakeholder to be sharing and managing their information in one place. We applied the agile methodology which allows the users to be involved in the entire processes of a website development. The results shown that since our community website located at www.pongyeangtravel.com has been launched, it has fifty registered tourist agencies that keep the contents up to date. During the first high season period, the website attracted five thousands pageviews, and the ReviewPongyeang Facebook page has around seven hundreds number of likes.

To better understand the characteristics and necessities of e-lab users, it is crucial to analyse the traffic of e-lab wiki website - an extensive set of online pages dedicated to experimental physics that serves as a support for the remote laboratory e-lab. The website’s traffic was carefully monitored during a 7 month period, covering variation of page views with time, most relevant pages, a demographic study of its users and the types of devices used to access the contents. The results showed a huge importance of non developed countries, specially India and Brazil, to the overall traffic. E-lab wiki has been accessed from more than 100 countries around the world, mainly through the use of mobile devices. Regarding accessed content, pages related to the determination of Planck’s constant composed the majority of the registered traffic.

Wikipedia is currently one of largest source of digital information. With billion page views per month worldwide, the free digital encyclopedia is based on collaborative editing, in which any user, even anonymously, can edit and include content. Although such content is always evaluated by several users daily and also by more experienced ones, vandalism is still common, especially on pages with sensitive information. In this paper we evaluated the vandalism found on pages of LGBT computer scientists, comparing them with scientists who are not part of this minority and analyzing the types of vandalism found. Thus, this paper analyzes how being part of a minority can influence relationships in the digital society.

Search Engine Optimization is a process by which parts of websites or webpage are modified and improved based on the recommendations of some popular search engines (Anon, 2018). The common way of doing this is running the website or specific webpage in the analyzer tool then it will recommend or suggest changes in the webpage. In today’s webpage optimization to implement the changes they need to open source code of the page and modify it. They also need some software like file transfer protocol software to upload changes. The process seems to be messy especially for the non-IT personnel who wants to improve and maintain the website like marketing people. In this case we are developing a tool that can help the optimization process very easy in each webpage of the website. The tool will be created with web based dashboard that can easily navigate and implement changes.

To ensure undisrupted web-based services, operators need to closely monitor various KPIs (Key Performance Indicator, such as CPU usages, network throughput, page views, number of online users, and etc), detect anomalies in them, and trigger timely troubleshooting or mitigation. There can be hundreds of thousands to even millions of KPIs to be monitored, thus operators need automatic anomaly detection approaches. However, neither traditional statistical approaches nor supervised ensemble approaches satisfy this requirement in practice when facing large number of KPIs. A state-of-art unsupervised approach Donut offering promising results, but it is not a sequential model thus cannot deal with the time information related anomalies. Thus, in this paper we propose Bagel, a robust and unsupervised anomaly detection algorithm for KPI that can handle time information related anomalies, using CVAE to incorporate time information and dropout layer to avoid overfitting. Our experiments using real data from Internet companies show that, compared to Donut, Bagel improves the anomaly detection best F1-score by 0.08 to 0.43.

The use of social networking sites have been growing dramatically. This phenomenon promotes most news agencies to use social media platforms rather than traditional media platforms for news releases. The main differences between social and traditional media platforms are “Real-time news release” and “cross-platform extensions Contents posted on social media platforms require speed and richness. Therefore, web news editors find it relatively difficult to write an article with good title in a short time that attracts audience to click in and read through. In this study, we analyzed selected “web news headlines with high page views” to identify catchy headlines and similarities of keywords. We used dimension reduction method as our primary technique to identify the best keyword combinations that help news editors to write attractive news for higher page views and more actual readers on the social media platforms.

"Well begun is half done". This proverb is especially true for a web product when it comes to creating a delightful and proactive user experience. This article describes our work in the last few years optimizing new user experience at LinkedIn, driven by application of data science and advanced analytical methods. Through mining the logs generated by new users in the past, we uncovered signals from their initial session that can predict their retention. We established the difference between creating models for predictions and creating models to inform product strategy. We found that persistent features, such as a user's number of connections, and having a confirmed channel of communication (email, phone or app), more strongly predict new user retention than most transient features such as how long they spend on the registration form or how many page views they have visited. We further constructed a true north metric (Quality Signup) to drive our Growth team towards the right focus as they iterated through multiple versions of new user onboarding flows. The strong positive correlation between the Quality Signup metric and long-term retention, as well as the positive impact we have seen on the product over the last two years, validate our strategy to drive product roadmap through data-informed metrics.

Click baits are primarily used by online content publishers. Their purpose is to allure readers to click on a link and subsequently visit other articles by the same publisher, in order to increase page views and ad revenue. Most of the time click baits are used for pointing to low quality articles or thin content. The user falls into the publishers’ trap due to a misleading or incomplete title or content exaggeration. A bait article link might also appear on social network shares or within the search engines result page, the presence of such a link in 3rd party web sites having a negative impact on user experience. Hence, it is important to properly identify and report them. In this paper we present an academic research browser extension meant to be used in the click bait reporting process. The extension offers to users the possibility to explicitly report a click bait, a series of details about the bait link being extracted and logged for further analysis. Based on all the gathered data, our goal is to obtain a community driven click bait samples database that may subsequently be used as an input for different supervised learning algorithms for click bait detection.

UC Browser as a service provider of mobile browser has market share of 28.51% in Indoensian market that drove it into the most popular mobile browser. Many people have done review of UC Browser and they have personal argument based on the methodology. The data on google play describe that user complained about the problems occured in UC Browser. This impacts on the decrease of UC Browser monthly page view in 2016. Therefore, it is necessary to research on the user acceptance factor that affect the usage of UC Browser. This research combine Task Technology Fit (TTF) and Unified Theory of Acceptance and Use of Technology (UTAUT) with construct Attitude. Data collection was done using online questionnaires. For generation Y task characteristics, technology characteristic, task-technoogy fit, task comfortance, performance expectation, effort expectation, social influence, and usage intention have an effect on UC Browser usage. For Y generation, performance expectation, facilitating condition, and user attitude of UC Browser are not the things that affect UC Browser usage. Based on the test results, the acceptance of UC Browser of Y generation using the model TTF and development UTAUT with Attitude construct showed that the model is able to explain 54.9% of the variance in the formation of UC Browser usage behavior.

Inbound marketing is the process of attracting the probable customers to a business before they have any intention to become customers. An effective method for inbound marketing is creation of a positive psychological business environment to attract the customers. A significant portion of traditional business environment is moving online and the new business environment is the company website. One of the major elements in online inbound marketing is the website address and the website logo, which are the first factors of brand personality that the visitor to the company website encounters when looking up the website in a search engine. In this paper, a framework for inbound e-marketing using visual and phonetic user experience analytics is proposed. The popular websites are studied and the relationship between website page views and the English phonetic construction of the website address and its logo are analyzed. For demonstrating the relationship between the website logo and name and its appeal to the customers, the proposed model is trained by a neural network. The proposed model is capable to predict website page views based on the company logo and the website address. The experimental results show that the proposed framework is capable of recommending the strategy for inbound marketing for online business and services with high accuracy.

Hollywood is a multi-billion dollar industry which releases more than a hundred films a year, with large variations in the budgets and box office grosses of the movies. Identifying which factors are important to a movie's profitability and subsequently predicting the success of a movie given its relevant parameters could save movie studios hundreds of millions of dollars a year. This paper analyses the efficiency of using multiple linear regression and Support Vector Machine Classification to predict the box-office success of movies, while analysing the influence of variables like trailer views, Wikipedia page views, critic ratings and time of release.

Many web-facing companies rely on online experimentation for evaluating users' reaction to new product features as part of the product development cycle. The existence of outliers often complicates the analysis of experiment results. Identifying outliers in online experiments is challenging as the distribution of web data such as page views do not satisfy normality assumption and are highly right-skewed. In this paper, we first illustrate the impact of outliers on experiment results and demonstrate the importance of outlier detection. As a solution, a generic method is necessary to handle outlier filtering across various types of web data. We introduce a statistical algorithm for detecting outliers regardless of the mechanism of their generation. Evaluation based on actual and simulated web data demonstrates that the proposed algorithm outperforms the other existing methods.

Nowaday, People can access information rapidly due to technology is higher than former times. Website is tool for present information of organization. The most important thing which consider is website content because of the content will explain everything which organization presented. The most of educational institution use website for presented their information to outside. However, educational institution website made people trust of them. Especially, Ranking of educational institution website will add value to the institution too. There are many ranking which ranked website of education for different purpose. Webometrics Ranking of World University is the ranking of education institution website which assessed accessibility of website content. Whereof, propose of the ranking is evaluate the accessibility of website which focus on visibility and activities. This article aims to evaluate the total number of backlinks, number of page view, duration of visit average times and bounce rate of Thailand University's website to finding relationship between number of backlinks, page view average, visit durations times, percentage of bounce rate and impact ranking of webometrics.

This paper describes the analysis and visualization of a Moodle student event log. A method for automatically generating pivot and time series cross section (TSCS) tables is proposed, and experimental results are summarized. The event log was accumulated in a face-to-face blended classroom with a large number of students using course materials uploaded on Moodle. For the accumulated event log, an experiment to automatically generate pivot and TSCS tables with the developed Excel macros was performed. As a result, both overall and individual viewpoints could be visualized for course material page views within classes and outside of class, viewing of all course materials and individual course materials, and viewing behavior of collected and individual students. The generated pivot and TSCS tables could be used to clearly identify students who did not open the course materials during class and those who opened the course materials later than instructed by the teacher. The results of the analysis in this paper can be used by teachers to look back on the progress of their lessons and provide suggestions for improvement.

Online social networks have a profound impact on our daily lives. Netizen is very concerned about how to get a better dissemination of message. Researchers usually use the SIR dissemination model, describing the basic model of the dissemination process, to explore the mechanism of message dissemination. In this paper, we used the SIR dissemination model to carry out the experiment based on the scale-free feature of the social network. We test our dissemination model from SIR on twitter network and compare the disseminating results under two main factors. Finally, we have got three results: 1, there is a phase transition in the message dissemination process. That is to say, all the messages that forwarded by a large number of users have jump of the number of page views. 2, The message content is more influential than the media. The probability of message being forwarded is more likely to affect its final dissemination than the popularity of the message senders. So, it is a good way to improve the quality of information if you want your message to be seen by as many people as possible. 3. The probability of meeage being forwarded has a threshold that affects its ability to spread. The probability above this threshold will easily get mass dissemination.

Nowadays, social networks are, in a technological level, one of the most important forms of communication, and are one of the most modern formats of connection between companies and their employees, as well, as a more active form of the companies to reach their customers. The online social network Facebook brings together approximately one billion users worldwide, according to the organization itself, and in Portugal, is the Website with the highest number of page views per month. Therefore, this article sought to verify how the gyms (was studied several pages gyms), behave in Online use of social networks, for communication between themselves and their employees, as with its members and future members. Through an online questionnaire directed to people who use and are familiar with Online Social Networks, sought to check what kind of Online Social Networks the respondents use, particularly Facebook, and if they use the same to search events in gyms that they attend. The work carried out has shown that people being familiar with Facebook, still do not use it to research information related to their physical activities.

Wikipedia is one of the most popular information platforms on the Internet. The user access pattern to Wikipedia pages depends on their relevance in the current worldwide social discourse. We use publically available statistics about the top-1000 most popular pages on each day to estimate the efficiency of caches for support of the platform. While the data volumes are moderate, the main goal of Wikipedia caches is to reduce access times for page views and edits. We study the impact of most popular pages on the achievable cache hit rate in comparison to Zipf request distributions and we include daily dynamics in popularity.

Search Engine Optimization (SEO) is important for websites to improve the rank for search results and get more page views requested by the user. This search engine ranks provide the better and optimized result to user, which will help them to view the popular page among the number of pages available in the search results. Apart from this search engine ranking, it also enables the websites to compete with other rival's website as each and every website owner expect to see their own website on the list before other's websites. This paper puts forward idea about SEO principles and basic strategies. It also expresses different techniques that are employed by search engines to improve its results. Also it presents the observation section, which gives the comparative analysis of SEO techniques.

Vast growth and the broad accessibility of data on the web have driven the surge of exploration movement in the zone of data retrieval on the Web. Topic and Trust are the important factors for data retrieval framework. As the extent of the web is huge, it is difficult to fulfill the user's need. To this end, the Web offers a rich site of data, which communicates through the hyperlinks. This paper explains the idea of improving PageRank, "Upgraded Page Rank with Topic utilizing Trust Component" which has the vast limit as compared with Conventional Page Rank Algorithm. In Page Rank algorithm, Page Rank score of a page ignores whether a page is relevant to a query or not, this becomes drawback and in the paper, a solution is presented by taking trust into consideration along with topical aspects.

This paper provides a web application design method, page file-oriented web application design with relevant concept and the corresponding definitions for the variety of web application development, technology and software engineering. We provide the page file flow chart which describes the relationships and parameters passing between every page of the website. It provides a standard technical specification for the website designing and development. This paper also gives mathematical definition of page file flow chart that can be further realizing the software package or plug-in to provide the support for visual website development and testing.

Clickbaits, in social media, are exaggerated headlines whose main motive is to mislead the reader to “click” on them. They create a nuisance in the online experience by creating a lure towards poor content. Online content creators are utilizing more of them to get increased page views and thereby more ad revenue without providing the backing content. This paper proposes a model for detection of clickbait by utilizing convolutional neural networks and presents a compiled clickbait corpus. We create a corpus using multiple social media platforms and utilize deep learning for learning features rather than undergoing the long and complex process of feature engineering. Our model achieves high performance in identification of clickbaits.

With the development of Internet technology, locating and identifying the page needed quickly from a sea of pages has become one of the most pressing and important demand. For the sake of acquiring video pages from mass web pages and analyzing their principal characteristics, this paper proposes the technology, which combines the form and content feature of the video page with topic web crawler of search engine technology, to identify video pages and video portals for the first time. To feed back to users about the results. So that they can access all video information on the Internet quickly. The result and analysis of environment shows that the algorithm is able to detect video pages with higher efficiency and accuracy.

This paper proposes to utilize a search engine as a social sensor whichis to be used for predicting market shares. More specifically, thispaper studies a task of comparing rates of concerns of those who searchfor Web pages among several companies which supply products, given aspecific products domain. In this paper, we measure concerns of thosewho search for Web pages through search engine suggests. Then, weanalyze whether rates of concerns of those who search for Web pages havecertain correlation with actual market share. Furthermore, as anintermediate statistics between the rates of concerns of those whosearch for Web pages and market share, we also examine correlationsbetween the page view statistics at the kakaku.com site and the othertwo statistics. The results of the analysis show that those threestatistics have certain correlations among each other. This paperfurther presents an empirical study on measuring an optimal correlationbetween the rates of concerns of those who search for Web pages andmarket share, as well as that between the rates of concerns of those whosearch for Web pages and the page view statistics at the kakaku.com site.

The paper proposes a new scheme named as WPP (Web Page Personalization) for effective web page recommendations. WPP consist of page hit count, total time spent in every link, number of downloads and link separation. Based on these parameters the personalization has been proposed. The system proposes a new implicit user feedback and event link access schemes for effective web page customization along with domain ontology. The proposed system builds a generalized inverted index framework for fast result finding and recommendation. The major goal of this paper is an approach of personalizing users' searches for achieve better search. To get a better item relevance estimation, the system uses the following parameters. (1). Event monitoring and user's click behaviors from Web search, (2). FLAME clustering.

A Uniform Resource Locator (URL) represents the address of a web page in World Wide Web (WWW). A URL grants access to a single web page on the WWW. Here in this paper the main focus is on the URLs addressing the same web page/same content. A web Page can have two or more URLs with which the web page can be accessed. These duplicate URLs can be a serious threat to the entire pipeline of internet searcher administration for indexing and creeping. I am presenting a novel algorithm for detecting canonization rules for normalizing URLs to the original single URL. Here a pattern recognition approach has been used for analyzing textual data. This approach benefits search engines from information about duplicate URLs to optimize the performance of search engine in terms of reduced cost and improved quality.

Page Views (PVs) are very crucial for search engines due to their close relationship to the revenue. When PVs change significantly, operators must be informed so that they can diagnose and fix the problem quickly, and prevent further loss. In reality, PVs can be counted in many ways (e.g., PVs originated from different ISPs), and different PVs are of different interest to operators (e.g., the PVs of a larger ISP is more important). As a result, different PVs often require different detection standards, or thresholds. However, attempts to tune a number of thresholds have been hampered by the cost of the manual effort involved. To address the above problem, we propose a practical framework, called PTL (practical threshold learning). Operators only need to provide a few simple labels about the detection results, then PTL will automatically tune the thresholds for different PVs. Using 4-month PVs from a global top search engine, our evaluation demonstrates that PTL can improve the accuracy of detection dramatically. More importantly, it introduces very little labeling overhead for operators. For example, when detecting the PVs of 103 ISPs, PTL can reduce the overall false negative rate from 96% to 9% using only 29 labels per week on average.

With the rise in popularity of using websites to distribute content to users, content creators needed to gain revenue from page views. To achieve this, they use third-party networks to distribute advertisements and perform analytics on users to customize their advertisements. These networks build up a profile of their users (e.g., navigation patterns, visit frequency, personal interest), which poses a significant privacy risk as this information can be sold to others or used in targeted advertising. In order for users to protect themselves, they need to install add-ons for their web browser which removes third-party content. This paper surveys some of the existing add-ons to block third-party content, and determine which add-ons are required to provide adequate protection. Further analytics are conducted to evaluate the effectiveness of those add-ons through the experiments.

Web page readability can be defined as the combination of reading comprehension, reading speed and user satisfaction. To improve the readability of a web page, content extraction and transformation techniques are used to present the main content to the reader in a more readable fashion. In this paper, we present the design of a readability tool which can be used as a web browser add-on. We aim at enhancing both the reading speed and comprehension. To achieve these goals, we extract and segment the main content into smaller coherent semantic units. These units are further augmented with text signals such as section headings, captions and page numbers in order to convey the text organization and the page visual structure; thus, enhancing the content comprehension. Our proposed system particularly suits constrained display devices such as mobile phones and PDAs.

This paper introduces DACP (Data Analyzer of Courser a Platform), a simple tool kit based on SQL commands and open source machines learning software WEKA to analyze the MOOC Courser a platform data including click stream, user information, lecture view, page view, video view, course assignment submission and grading, quiz submission and grading, course grades and so on. It includes five modules to implement the data transformation, modification, updating and analysis. As a test bed, the authors used this toolkit to analyze the data of one MOOC from Peking University in details. The descriptive and correlative analysis result shows that the learners' participation in all course activities is the determinant factor to their learning achievement in a MOOC.

Network arousing more and more attention at present, it is important to grasp the hot topic for news editing and public option surveying. For regional BBS forums, the hot topic is issued in the way of forum posts by individuals, instead of by official media. So the information is various, and the hot topic is hard to be found out. This paper introduced a regional BBS-oriented hot topic a tracking system based on the page views and replies. By processing the page views and replies extracted from the web pages collected and combining with the cluster analysis, the system could get rid of the interference effectively, and detects the hot topic correctly.

World is full of information. The World Wide Web serves as major source of getting such information. Due to the changing nature of web plenty of web pages are deleted and added newly. Every time a surfer searches web using the search engine, data should be fresh and relevant. Retrieving efficient, relevant and meaningful information from these large sources of information is very challenging job. Due to the large size of web, corresponding to any query made by a user number of pages are being retrieved so the result should be ordered in the manner that most relevant web pages is on top of the list. In order to get most relevant pages at top and reduce the problem of theme drift we propose a hybrid approach of Enhanced-Ratio Rank and Page level keyword algorithms.

This paper is focused on computing importance of a web page in an efficient way. Web page ranking is an essential factor in web search. Many modules and algorithms have been proposed using different resources with different assumptions. The algorithms proposed include Page Rank, Browse Rank, Browse Rank Plus, HITS and many more. Page Rank focuses on ranking a page based on the number of inlinks and outlinks to a page. Whereas Browse Rank focuses on ranking the page based on the value it provides to the user. Several other algorithms have been proposed since, that focuses only on one or two particular factors. This paper proposes ranking a page based on multiple factors that includes reachability, value and user feedback. The major aim is to rank a web page based on these three crucial factors rather than considering one or two factors taken into account by existing methodologies. Every user has a different and unique background and a particular aim when searching for information on the Web. Web search personalization is mainly aimed at tailoring search results to a specific user based on that user's interests and preferences. Major challenges that effective personalized search is affected with includes accurately identifying the user context and organizing the information in such a way that it matches the particular context. An effective mechanism is employed to personalize the search and also to rank the page based on multiple factors.

Web mining is a technique through which various analyses can be done so that it can be used for a variety of applications such as page ranking. Page ranking denotes the popularity or usage of web pages over internet. Although there are various page ranking algorithms implemented, but HITS and Google Page ranking are the most popular raking algorithms. This technique provides efficient ranking of web pages and provides less computational time, but here in this paper the algorithm is enhanced on the basis of response time of each web page. The idea is to compute access time of each web page and then compute average response time on the basis of which ranking of web pages can be predicted. The technique implemented here provides efficient page ranking as well as provides better goodness factor.

This article gives an overview of the currently available literature on web page ranking algorithm using machine learning. Web page ranking algorithm, a well-known approach to rank the web pages available on cyber world. It helps us to know -- how the search engine exactly works and how a machine learn itself while giving priority to the page that which page is important to successfully fulfills the user query need and which page is worth less. Machine learning approach also helps us in understanding the complex part of page priority criteria in most popular search engines like Google, yahoo, AltaVista, dog pile and many more search engines like that. Page ranking mainly unrevealed the structure of web. This article gives an overview of available literature in the field of web page ranking algorithm and it also highlights the main point based on machine leaning approach in web page ranking algorithm.

Aggregate answer ratings serve as a metric of collective intelligence in social QaA communities. The patterns by which participants in a social QaA community rate and recommend answers are analyzed through the lens of first-mover advantage, to address the question of whether the first answer posted has a ratings advantage over those subsequently submitted. As part of a long-term participant observation, ratings for answers submitted to the Answer bag social QaA site were compared by order of submission and normalized for page views and answer quality. The results suggest that the first-submitted answer consistently accumulates roughly 17% more rating points than the second answer submitted, and that the rating points of each subsequent answer tend to decline. Social factors influencing rating activity and implications for interpreting future social QaA data are discussed.

In the current diversity and complexity of the network information environment, the technology of web page sensitive keywords detection is an important and immediate way to manage public opinion online. We propose a system for web page sensitive keywords detection. This system can detect sensitive keywords in the web pages timely and effectively. And it will mark the position of the keywords in web pages and writes the related information such as detection results and detection time to the detection result table in order to help managers of websites to take measures to filter sensitive keywords in time.

Nowadays the usage of mobile phones is widely spread in our daily life. We use mobile phones as a camera, radio, music player, and even an internet browser. As most Web pages were originally designed for desktop computers with large screens, viewing them on smaller displays involves a number of horizontal and vertical page scrolling. To save mobile Web search fatigue caused by repeated scrolling, we investigate the automatic Web page scrolling problem based on two observations. First, every web page has many different parts that do not have the equal importance to an end user, and the user is often interested in a certain part of the Web page. Second, the ease of use of text-entry in mobile phones compare to the desktop computers', users usually prefer to search the Web just once and get the needed answer. Compared to the existing efforts on page layout modification and content splitting for easy page navigation on mobile displays, we present a simple yet effective approach of automatic page scrolling for mobile Web search, while keeping the original Web page content keeps its integrity and hence, preventing any loss of information. We work with the Document Object Model (DOM) of the clicked page by user, compute the relevance of each paragraph of the Web page based on the tf*idf (term frequency*inverse document frequency) values of user's search keywords occurring in that paragraph. The focus of the browser will be automatically scrolled to the most relevant one. Our user study shows that the proposed approach can achieve 96.47% scrolling accuracy under one search keyword, and 94.78% under multiple search keywords, while the time spending in computing the most important part does not vary much from the number of search keywords. The users can save up to 1.5 sec in searching and finding the needed information compare to the best case of our user study.