1,721,012 research outputs found
BSc-project: Geautomatiseerde webstatistiekanalyse & website-prestatie-indicatie
No full textContent Power biedt hun klanten een Content Management Systeem (CMS) aan waarmee zij gemakkelijk content op hun website kunnen aanpassen of toevoegen. Voor dit Bachelor-eindproject hebben wij onderzoek gedaan naar geautomatiseerde webstatistiekanalyse en website-prestatie-indicatie voor de websites van klanten van Content Power. Hiervoor hebben wij literatuuronderzoek gedaan, interviews gehouden en uiteindelijk een proof-of-concept gebouwd. Deze proof-of-concept bestaat uit een aanbevelingen-dashboard als module binnen het CMS van Content Power. Websitebeheerders krijgen met behulp van een expertsysteem gegenereerde aanbevelingen te zien met als doel meer conversies en search engine-optimalisatie. Tevens worden de beheerders in staat gesteld de prestaties van hun website te volgen door middel van door hun instelbare doelstellingen. Aan eenvoud wordt een hoge prioriteit gegeven alsmede het voorzien van de module van een "actionable interface". Vanwege de complexiteit van dit project en de gelimiteerde tijd is er gekozen voor een proof-of-concept. Daarbij is veel aandacht besteed aan het uitbreidbaar maken van ons systeem, zodat hier later door Content Power op verder gebouwd kan worden.BSc Technische InformaticaIntelligent SystemsElectrical Engineering, Mathematics and Computer Scienc
Where will you comment next? Exploiting comments for personalized recommendations
No full textSince the advent of Web 2.0, users have not only increasingly created content, but also contributed reactions to content in the form of comments. Comments are challenging to analyze due to their short lengths and informal style, meaning that any individual comment provides very little data to work with and is highly variable. However, comments capture innate and an explicit opinion of a user that makes it invaluable towards personalization. In this work, we explore the possibilities of exploiting comments towards the end of personalized recommendations. Over the course of this work, we investigate the particular domain of news recommendation and report our findings through use of different recommenders evaluated offline. Our contributions include an evaluation strategy that allows for simulation of recommenders offline, a simplistic hybrid filtering technique that exploits the advantages of its root recommenders and various findings related to news recommendation in general. We perform a preliminary study into investigating whether users maybe attributed by comments they make and find that they are indeed attributable if the right features are considered. Utilizing the property of authorship attribution through comments, we achieve user-user similarity that ultimately aids in delivering recommendations. We find that freshness is an important aspect in news recommendation and therefore for the design of our recommender we build upon the freshness aspect while also achieving personalization by exploiting content, user-user similarity and the user neighbourhood.Multimedia ComputingIntelligent SystemsElectrical Engineering, Mathematics and Computer Scienc
Exploiting Embedding in Content-Based Recommender systems
No full textXING is a leading career-oriented social networking site in Europe, which usually recommend job ads to their customers. One of the widely used methods in Recomender Systems is content-based filtering, which analyzes the description of item characteristics and the user profile illustrating user's preferences. Due to the sparsity of its dataset, i.e. many job postings are rarely interacted with, XING has been using content-based recommender system to promote the quality of the recommendations. Recent word embedding technique learns semantically meaningful representations for words from co-occurrence in sentences, which enables the effective comparison between words. Based on the Word2Vec technique, XING represents job postings by the average embedding over words they contain. This study explores three alternative methods to represent job postings for the task of recommending jobs to users. In the first experiment, we explore whether the use of a subset of words is more effective to represent the job postings. In the second experiment, instead of averaging over word embeddings, we directly learn document embeddings using Paragraph2Vec. And finally, the third experiment uses Word Mover's Distance to estimate the similarity between job postings. Our experiments show that the embeddings that are learned with Paragraph2Vec result in a better estimation of which job postings are similar, but only when high-dimensional settings are used. The Word Mover's Distance algorithm is computationally expensive, therefore we use existing lower-bounds that allowed us to complete a small-scale experiment within the available time. The results indicate that Word Mover's Distance is not as effective as the average over word embeddings and Paragraph2Vec. In the final part of this thesis, we present the Link2Vec, a novel item representation method based on Word2Vec, which learns semantic representations for items based on the context surrounding the hyperlinks that refer to the item, e.g. hyperlinks to the item's Wikipedia page. Our experiments show that the effectiveness of the embeddings learned with Link2Vec improves with the amount of training data. For the evaluation on the MovieLens dataset, we only obtained a limited set of hyperlinks, which resulted in results that approximate a baseline that uses the average over word embeddings.Electrical Engineering, Mathematics and Computer ScienceIntelligent System
Image-Based News Recommendation
No full textElectrical Engineering, Mathematics and Computer ScienceIntelligent SystemsMultimedia Computin
Exploiting Local Semantic Concepts for Flooding-related Social Image Classification
Full text linkContains fulltext :
199472pub.pdf (Publisher’s version ) (Open Access)MediaEval 2018 Workshop, 29 oktober 201
Do The Math
No full textDo The Math is a game, developed in a BSc programme, by two students. The goal is to enable people to practice their math skills. The main focus of the project was the scalability of the backend systems.Game technologySoftware Computer TechnologyElectrical Engineering, Mathematics and Computer Scienc
Ranking and Context-awareness in Recommender Systems
No full textIn this thesis we report the results of our research on recommender systems, which addresses some of the critical scientific challenges that still remain open in this domain. Collaborative filtering (CF) is the most common technique of predicting the interests of a user by collecting preference information from many users. In order to determine which items from a collection may be favored by individual users, conventional CF approaches take the ratings previously assigned to items by a target user and use them together with ratings of users with similar preferences to predict the ratings of yet-unseen items. Then, items are recommended in a descending order according to their predicted ratings. While CF has been investigated and improved extensively over the past years, there is still room for substantial improvement. In this thesis we focus on improvement of two critical aspects of CF, namely ranking and context-awareness of the recommendations. In addition, we analyze new developments in the field of collaborative recommendation and elaborate on the challenges related to the evolution of recommender systems and their increasing impact in the future. Based on this analysis, we make recommendations for future research directions in this field.Intelligent SystemsElectrical Engineering, Mathematics and Computer Scienc
Going Beyond Counting First Authors in Author Co-citation Analysis
Full text linkThe present study examines one of the fundamental aspects of author co-citation analysis (ACA) - the way co-citation
counts are defined. Co-citation counting provides the data on which all subsequent statistical analyses and mappings
are based, and we compare ACA results based on two different types of co-citation counting - the traditional type that
only counts the first one among a cited work's authors on the one hand and a non-traditional type that takes into
account the first 5 authors of a cited work on the other hand. Results indicate that the picture produced through this non-traditional author co-citation counting contains more coherent author groups and is therefore considerably clearer. However, this picture represents fewer specialties in the research field being studied than that produced through the traditional first-author co-citation counting when the same number of top-ranked authors is selected and analyzed. Reasons for these effects are discussed
Projectcampus grading and review
No full textThis report describes, analyzes and evaluates the process around the design of the grading and review module and its development for the online learning environment Projectcampus. The module was developed as a bachelor’s project. This project is the last part of the bachelor phase of the computer science study of the Delft University of Technology. The project entails developing a software product for an existing company while following normal software development procedures. The goals of this project are to teach students to work in a team of colleagues, in a real life company; to choose and follow a development method in cooperation with the client, and manage the products and processes that come with it; to determine quality requirements and assess them; and present and explain the results of the project. In this project, three students from the Delft University of Technology have developed a module for the existing Projectcampus system of Shareworks.Technische InformaticaComputer ScienceElectrical Engineering, Mathematics and Computer Scienc
- …
