EXTRACTING ATTRIBUTES AND THEIR VALUES FROM WEB PAGES

In this paper, we propose a novel method for extracting information from HTML tables with similar contents but with a different structure. We aim to integrate multiple HTML tables into a single table for retrieval of information containing in various Web pages. The method is designed by extending tree-structured LSTM, the neural network for tree-structured data, in order to extract information that is both linguistic and structural information of HTML data. We evaluate the proposed method through experiments using real data published on the WWW.

The usability of an interface is a fundamental issue to elucidate. Many researchers argued that many usability results and recommendations lack empirical and experimental data. In this research, the usability of the web pages is evaluated using several carefully selected statistical models. Universities web pages are chosen as subjects for this work for ease of comparison and ease of collecting data. A series of experiments has been conducted to investigate into the usability and design of the universities web pages. Prototype web pages have been developed according to the structured methodologies of web pages design and usability. Universities web pages were evaluated together with the prototype web pages using a questionnaire which was designed according to the Human Computer Interactions (HCI) heuristics. Nine (users) respondents’ variables and 14 web pages variables (items) were studied. Stringent statistical analysis was adopted to extract the required information to form the data acquired, and augmented interpretation of the statistical results was followed. The results showed that the analysis of variance (ANOVA) procedure showed there were significant differences among the universities web pages regarding most of the 23 items studied. Duncan Multiple Range Test (DMRT) showed that the prototype usability performed significantly better regarding most of the items. The correlation analysis showed significant positive and negative correlations between many items. The regression analysis revealed that the most significant factors (items) that contributed to the best model of the universities web pages design and usability were: multimedia in the web pages, the web pages icons (alone) organisation and design, and graphics attractiveness. The results showed some of the limitations of some heuristics used in conventional interface systems design and proposed some additional heuristics in web pages design and usability.

In 2020, Most Filipinos are using the internet due to COVID-19 pandemic lockdowns. The internet is not limited to adults and children might be exposed to online adult content and abuse. The Philippine Internet Service providers fail to capture pornographic web pages that are not for child viewing. A Web Page classifier would help in detecting and classifying web pages. In this study, a total of 12000 web pages with adult content and academic web pages were collected using scrapy and existing datasets from DMOZ were used to create a Support Vector Machine (SVM) multi-class classifier. To improve the accuracy of the SVM model, data preprocessing was performed to remove noisy and irrelevant data from the dataset. The text in the web pages was used to train the SVM classifier by using Term Frequency and Inverse Document Frequency, Count vectorizer, and Word2vec Skip-gram embedding with TF-IDF as a multiplier. A series of experiments were conducted using multiple word embedding techniques. The SVM model built using word2vec with TF-IDF multiplier outperforms the SVM model built using TF-IDF and Count Vectorizer. The word embedding generated using word2vec was generated with a window size of 9 and a vector dimension of 900. The SVM model built using word2vec shows an S6% accuracy. The SMV model is deployed in the Django framework and a chrome plugin was created to use the SVM model using REST API.

During the MEGAPIE Integral Test (MIT) of the IMW liquid metal (LM) target for a spallation neutron source, infrared thermography (IRT) has been used to investigate the liquid lead-bismuth eutectic (LBE) cooling of the proton beam entry window of the target. The IRT investigations have been performed on the real MEGAPIE target on the MIT stand (MITS) at Paul Scherrer Institute (PSI) in Switzerland, few months before of the proton irradiation period of the target. In meantime the target has been successfully irradiated (from August to December 2006) at the PSI Spallation Neutron Source (SINQ) facility; for more details see web page http://megapie.web.psi.ch/.<br />The Goals of the MIT thermo-hydraulics investigations of the liquid LBE "flow in flow" window cooling configuration (ca. 1.25 m/s speedy by-pass jet flow into the ca. 0.33 m/s main flow), have bifocal perspective.<br />• On one hand the goal was visualization for an observation of changes of the target cooling field pattern of the by-pass jet flow, i.e. the qualitative investigation of cooling behavior in terms of the sufficient geometrical covering of the proton beam irradiation footprint area.<br />•On second hand the goal was determination of local convection heat transfer coefficient (HTC) on the steel wall of the proton beam entry window area of the MEGAPIE target, i.e. the quantitative values and distribution of a cooling.<br />For the qualitative visualization of the real, "in situ", target window cooling we have take advantage of slightly higher temperature of the by-pass jet flow LBE streaming (ca. 4°K) and we have performed so called "MIT-Warm-Jet" experiment series. For the quantitative determination of HTC instead of the real target window we have used specimen sensor dish and performed so called "MIT-KILOPIE" experiment series using the two dimensional and dynamic infrared thermography (2DD-IRT) method, which has been developed and tested in 2005 at PSI; reported during Thermosense XVIII in 2006.<br />The results of measurements are presented in form of IRT thermograms or thermogram sequences which are extracted from the raw temperature field measurements. The in December 2006 successfully finished irradiation period of MEGAPIE project and the integrity of the whole target have shown and proved the predicted sufficiency of the window cooling.

This paper proposes a neural network-based method for generating compact answers to open-domain why-questions (e.g., "Why was Mr. Trump elected as the president of the US?"). Unlike factoid question answering methods that provide short text spans as answers, existing work for why-question answering have aimed at answering questions by retrieving relatively long text passages, each of which often consists of several sentences, from a text archive. While the actual answer to a why-question may be expressed over several consecutive sentences, these often contain redundant and/or unrelated parts. Such answers would not be suitable for spoken dialog systems and smart speakers such as Amazon Echo, which receive much attention in these days. In this work, we aim at generating non-redundant compact answers to why-questions from answer passages retrieved from a very large web data corpora (4 billion web pages) by an already existing open-domain why-question answering system, using a novel neural network obtained by extending existing summarization methods. We also automatically generate training data using a large number of causal relations automatically extracted from 4 billion web pages by an existing supervised causality recognizer. The data is used to train our neural network, together with manually created training data. Through a series of experiments, we show that both our novel neural network and auto-generated training data improve the quality of the generated answers both in ROUGE score and in a subjective evaluation.

Problem solving methods, inspired biologically or by nature, are gaining in importance. Their potential for devising new, but also for increasing efficiency of some well-known methods for solving different kinds of problems, is becoming more and more obvious. The behaviour of social insects provides a lot of inspiration. We took inspiration from honey bees, in particular the way they behave in a hive. We proposed a modified model devised for searching parts of the web online. In a series of experiments, we attempted to follow a developing story with the aim to automatically find documents (web pages) bringing news on the story. This service, working as a beehive, delivered documents as expected.Keywordsbee hive modelbee swarm intelligenceweb searchon-line web crawlingdeveloping story

This paper proposes a method for retrieving Web pages according to objects described in them. To achieve that goal, ontologies extracted from HTML tables are used as queries. The system retrieves Web pages containing the type of objects described by a given ontology. We propose a simple and efficient algorithm for this task and show its performance on real-world Web sites.KeywordsInformation RetrievalVector Space ModelBlock SequenceRelevance ScoreQuery WordThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Suppose that we wish to know a group of proteins responsible for a certain cellular biological process. Here we propose to infer such a protein complex from a protein-protein interaction network by using a class of algorithm, which has originally been developed to achieve web page ranking that reflects user’s personal interest or context. The inference of proteins responsible for a given biological process, namely, personalized ranking of proteins is whereby performed in analogy with personalized ranking of web pages. Searching for the best approach to personalized protein ranking, we carry out a series of experiments to compare the performance between two major personalized ranking methods: the personalized PageRank algorithm and the continuous-attractor dynamics algorithm, both applied to a yeast protein-protein interaction network. Results of these comparison experiments suggest that the continuous-attractor dynamics algorithm is the most efficient for personalized protein ranking.KeywordsBioinformaticsInteractomeProtein-protein interaction networkPersonalized PageRank algorithmContinuous-attractor dynamics

Growing concern over digital privacy has led to the widespread use of tracking restriction tools, such as ad blockers, Virtual Private Networks (VPN), and privacy-focused web browsers. All major browser vendors have also deprecated, or plan to deprecate, third-party cookies to reduce tracking. Despite these efforts, advertising companies continuously innovate to overcome these restrictions. Recently, advertising platforms, like Meta, have been promoting server-side tracking solutions to bypass traditional browser-based tracking restrictions. This paper explores how server-side tracking technologies can link website visitors with their user accounts on Meta products. The goal is to assess the effectiveness and accuracy of employing this technology, as well as the effect of tracking restrictions on online tracking. Our methodology involves a series of experiments where we integrate Meta's client-side tracker (the Meta Pixel) and server-side technology (the Conversions API) on different web pages. We then drive traffic to these pages and evaluate the success rate of linking website visitors to their profiles on Meta products. Our findings show that Meta's server-side technology can match between 34% and 51% of website visitors to user profiles on Meta products using basic information like the visitor's IP address, user agent, and location data. This is comparable to Pixel-based user matching in optimal conditions (i.e., in the absence of tracking restrictions), which links between 42% and 61% of user profiles. Nevertheless, we see a considerable difference in accuracy: while the Pixel-based tracking achieves 100% accuracy, less than 65% of the profiles matched by server-side tracking are accurate.

Multi-label problems are challenging because each instance may be associated with an unknown number of categories, and the relationship among the categories is not always known. A large amount of data is necessary to infer the required information regarding the categories, but these data are normally available only in small batches and distributed over a period of time. In this work, multi-label problems are tackled using an incremental neural network known as the evolving Probabilistic Neural Network (ePNN). This neural network is capable of continuous learning while maintaining a reduced architecture, so that it can always receive training data when available with no drastic growth of its structure. We carried out a series of experiments on web page data sets and compared the performance of ePNN to that of other multi-label categorizers. On average, ePNN outperformed the other categorizers in four out of five metrics used for evaluation, and the structure of ePNN was less complex than that of the other algorithms evaluated.

The evolving Web has seen ever-growing use of structured data, thanks to the way it enhances information authoring, querying, visualization and sharing. To date, however, most structured data authoring and management tools have been oriented towards programmers and Web developers. End users have been left behind, unable to leverage structured data for information management and communication as well as professionals. In this paper, I will argue that many of the benefits of structured data management can be provided to end users as well. I will describe an approach and tools that allow end users to define their own schemas (without knowing what a schema is), manage data and author (not program) interactive Web visualizations of that data using the Web tools with which they are already familiar, such as plain Web pages, blogs, wikis and WYSIWYG document editors. I will describe our experience deploying these tools and some lessons relevant to their future evolution.

New interaction experiences are fundamentally changing the way we interact with the web. Emerging touch-based devices and a variety of web-connected appliances represents challenges that prevents the seamless reach of web resources originally tailored for the standard browser experience. This paper explores how web pages can be re-purposed and become interactive presentations that effectively supports communication in scenarios such as digital signage and other presentation use cases. We will cover the TagVisor project which is a JavaScript run-time that uses modern animation effects and provides an HTML5 extension approach to support the authoring of visual narratives using plain web pages.

In recent years, web-based appointment platforms develop rapidly, and many of them provide physician recommendation services to patients. Considering patients’ heterogeneity in both illness and behavior, it is a great challenge to deliver personalized recommendations. Motivated by this healthcare application, this study incorporates patient choices and focuses on optimizing real-time personalized recommendation of physician assortment with limited resources, in order to satisfy patients’ varying demands and improve the resource allocation efficiency. This work considers not only the influence of physician assortment to patient choice but also how physicians in the assortment are displayed on the webpage, i.e., the order of physicians. We adopt the location-based two-stage choice model to capture the patient behavior, in which patients randomly view the top physicians and then choose one among them. The recommendation problem is studied in both static and dynamic environments. In the static environment, we ignore resource capacities and optimize the recommendation of physician assortment as well as the displayed ranking. We propose a heuristic algorithm SORT for this static version of the problem and prove the lower bound of algorithm performance, along with the numerical performance validation. In the dynamic environment, we optimize the physician recommendations for a sequence of randomly arriving patients considering patients’ heterogeneity in both matching degrees and choice probabilities. We propose a dynamic algorithm Adjust-exponential inventory balancing (Adjust-EIB) by incorporating our static algorithm SORT in the improved existing algorithm, which makes recommendation decisions based on the real-time remaining resources. We conduct a series of numerical experiments to compare our algorithm with several benchmarks. The numerical results show that Adjust-EIB outperforms the benchmark algorithms, especially in congested systems. We also conduct a case study with real-world data and verify the capability of our algorithm in improving real-world system efficiency. Note to Practitioners —This work is motivated by the increasing popularity of web-based appointment platforms. Considering the fact of limited physician resources, we address the issue of how to recommend physicians for patients in a personalized and real-time way on web-based appointment platform, in order to improve the matching degree between physicians and patients as well as the resource allocation efficiency. To the best of our knowledge, we propose one of the initial methods to recommend personalized physician rankings in dynamic environments with limited resources. We construct a specific model and propose algorithms to make physician recommendation decision, which performs better than other benchmark approaches based on the numerical analysis, and the case study with real-world system data indicates our method’s capability of solving the practical problem. Our method is robust enough for the reason that the model allows arbitrary arrival pattern. The model and methods are specifically designed for the web-based appointment platforms, but it can also be easily applied in other application scenarios based on visual web pages, such as e-commerce.

Aim/Purpose: This paper describes an empirical investigation on how consumer behavior is influenced by the volume of content on a commercial landing page -- a stand-alone web page designed to collect user data (in this case the user’s e-mail address), a behavior called “conversion.” Background: Content is a term commonly used to describe the information made available by a website or other electronic medium. A pertinent debate among scholars and practitioners relate to information volume and consumer behavior: do more details elicit engagement and compliance, operationalized through conversions, or the other way around? Methodology: A pilot study (n= 535) was conducted in ‎real-world commercial setting, followed by a series of large-scale online experiments (n= 27,083). Both studies employed a between-group design: Two variations of landing pages, long and short, were created based on various behavioral theories. User traffic to the pages was generated using online advertising and randomized between the pages (A/B testing). Contribution: This research contributes to the body of knowledge on the antecedents and outcomes of online commercial interaction, focusing on content as a determinant of consumer decision-making and behavior. Findings: The observed results indicate a negative correlation between content volume and users’ conversions. The shorter pages had significantly higher conversion rates, across locations and time. Findings suggest that content play a significant role in online decision making. They also contradict prior research on trust, persuasion, and security. Recommendations for Practitioners: At a practical level, results can inform practitioners on the importance of content in online commerce. They provide an empirical support to design and content strategy considerations, specifically the use of elaboration in commercial web pages. Recommendation for Researchers: At the theoretical level, this research advances the body of knowledge on the paradoxical relationship between the increased level of information and online decision-making and indicates that contrary to earlier work, not all persuasion theories‎ are ‎effective online. Impact on Society: Understanding how information drive behavior has implications in many domains (civic engagement, health, education and more). This has relevance to system design and public communication in both online and offline contexts, suggesting social value. Future Research: ‎Using this research as a starting point, future research can examine the impact of content in other contexts, as well as other behavioral drivers (such as demographic data). This can lead to theoretical, methodological and practical recommendations.

The usage of PDA and mobile devices has dramatically increased recently. However, mobile devices and PDA devices have a limited screen size, which makes it frustrating to browse tabular data on mobile devices since users have to frequently scroll up and down to find the information of interest. This paper presents an efficient means to present HTML-based tables on mobile devices. Based on the column and row headers, the authors adapt a HTML-based Web table into two adaptive styles. The first style displays all information of a table into a single narrow page to avoid horizontal scrolling; and the second style distributes information to different sub-pages, each of which approximately occupies the whole mobile screen, and thus eliminates scrolling. The approach is empirically evaluated using a controlled experiment. The main conclusions derived from the empirical study are: (1) the adaptive layout styles improves the browsing efficiency for individual subjects as compared to HTML web page style, (2) the single narrow adaptive layout resulted in the improved browsing efficiency compared to the multi-page adaptive layout for one-dimensional HTML web page tables, and (3) the multi-page adaptive layout was more efficient than the single narrow adaptive layout for two-dimensional HTML tables.