Archived Past Project Ideas

Design Guidelines for Crowdsourcing Emotion Annotations [Archived]

Duration: One semester
Goals: Develop guidelines for designing a crowdsourcing task for textual emotion annotation.
Responsible TA: Valentina Sintsova
Student Name: Not assigned yet
Keywords: Emotion Annotation, Crowdsourcing, Design Guidelines, Design Recommendations, Task Design
Abstract:

Annotation of emotions in collected text documents is an essential step for developing the reliable emotion classification system. It can serve as the ground-truth for training and/or testing emotion recognition models. Current crowdsourcing approaches involve annotation of text documents for emotions and emotion indicators. Yet, subjectivity in understanding of the emotion concept and lack of clear instructions may result in high inconsistency of collected data. There is a need to elaborate the appropriate design guidelines in order to ensure the collection high-quality annotations.

The responsible student will:

  • Survey the existent approaches to crowdsourcing sentiment and emotion annotation of text, as well as the method of their qualitative and quantitative evaluation
  • Analyze data collected from the previous crowdsourcing experiments on emotion annotation
  • Based on that analysis, derive design guidelines for crowdsourcing tasks on textual emotion annotation
Related Skills: Strong analytical skills; Statistical analysis; Interest in Crowdsourcing, Qualitative Research, and Human-Computer Interaction Design
Suitable for: Bachelor or Master Student

Predicting Influence of Online Reviews with Emotion Extraction [Archived]

Duration: One semester (can be adapted to both bachelor and master students)
Goals: Improve our prediction capability of influence among reviews from e-Commerce websites, and our understanding of human behavior in this setting.
Responsible TA: Lionel Martin  and Pearl Pu
Keywords: emotion recognition, influence detection, prediction, social media analysis, natural language processing
Abstract:

Reviews keep playing an increasingly important role in the decision process of buying products and booking hotels. However, the large amount of available information can be confusing to users. A more succinct interface, gathering only the most helpful reviews, can reduce information processing time and save effort. To create such an interface in real time, we need reliable prediction algorithms to classify and predict new reviews which have not been voted yet but are potentially helpful. So far such helpfulness prediction algorithms have benefited from structural aspects, such as the length of review or its readability score. Since emotional words are at the heart of our written communication and are powerful to trigger listeners’ attention, we believe that they can serve as important parameters for predicting helpfulness of review texts. This is an excellent opportunity to develop a combined approach of Machine Learning and Natural Language Processing to improve the prediction of these influential reviews and provide a light interface to users.

The responsible student will:

  1. Familiarize with existing techniques in influence prediction and emotion extraction
  2. Discover and try the existing framework for prediction
  3. Develop new learning mechanisms to extract emotions from review texts
  4. Understand the features of interest in the prediction process
Related Skills: Programming skills (Python, Matlab or similar for prediction); Interest in Machine Learning, Natural Language Processing, and/or Text Mining; French or English mother tongue a plus

Persuasive technology in Group Lifestyle Recommender Systems [Archived]

Duration: One semester
Goals Persuasive technology aims to change people’s behavior or attitude through persuasion and social influence, but not coercion. Particularly, social influence has been identified as an effective component in motivating users’ behavior change in lifestyle. This project aims to develop a group lifestyle recommender system by applying persuasive technology based on a group recommender system called GroupFun.
Responsible TA: Yu Chen  and Pearl Pu
Keywords: Persuasive technology, social influence, behaviour change
Abstract:

The student will develop a group lifestyle recommender system based on a group music recommender system called GroupFun and conduct user studies to evaluate the system.

More concretely, the responsible student will perform the following tasks:

  • Qualitative research: including 1) conduct interview with real users to identify their goals, needs and requirement and 2) survey competitor products, relevant studies and identify novel approaches.
  • Prototype: Design and implement the lifestyle recommender using social influence based on qualitative research.
  • User study: Evaluate the system and persuasion technology with real users.
Related Skills: Java. Mobile app development experience  is a plus.

Tomorrow’s music player [Archived]

tomorrowsmusic

Title: Tomorrow’s music player
Duration: One Semester
Responsible TA: Nicolas Jones
Goals: The goal of this project is to design a new visual interface for an existing music recommender, Last.fm.
Solution: Extend the jlfm java library and create a music player (respecting the audioscrobbler protocol)
Keywords: audioscrobbler. interface, java, last.fm, library, middleware, mp3, protocol, streaming, visualisation
Abstract:

When the first music recommender systems appeared a few years ago, Last.fm and Pandora set the trend as for what the interface should look like. Since then, many new recommenders and social music players have emerged, but little have presented an innovative interface, preferring to copy the industry’s current standard. This project aims at introducing some novelty in this process by removing some buttons which can be replaced by analysing a user’s behavioural pattern and inferring adequate actions from his implicit preferences. This reduction of visual items in the interface brings up some place for new elements, such as visual representations of social similarity amongst users, tag clouds, and more. The advantage of such a system is to use the recommendation power of an accepted and popular system, but providing an innovative interface, resolving many of the classic user interaction issues of dynamic music recommender systems Current Java libraries (from the Jlfm project) provide basic access to Last.fm’s platform and allow to capture any stream of music. They can easily serve as a feed for a new interface. The interface could be written in Java, but could also be proposed as a web-page interacting with a server-instance of the existing Java libraries. The main steps of this project would be to first extend the communication means with Last.fm. creating a new music player (software or website) before working on additional features.

Required Skills
  • Good software engeneering skills (java or scala and possibly xhtml/css)
  • Basic HCI knowledge
  • Creativity and interest in music

iTunes automatic music classifier & scribbler [Archived]

ituplug

Title: iTunes automatic music classifier & scribbler
Duration: One Semester
Responsible TA: Nicolas Jones
Goals: Classify music (and, possibly, recommend better music or generate automatic playlists)
Keywords: iTunes, preference elicitation
Abstract:

iTunes already incorporates a rating mechanisms, a genre classification, a playlists categorisation, and a play-count log. Unfortunately, this information is very rarely used by iTunes to represent data in a semi-automatic way. Furthermore, a majority of users don’t use these at all, and the small subset of those who do spend a lot of time with it. We propose to write a plugin who will rely on the users’ implicit preferences (behaviour) to rate the songs, classify them by similarity in automated playlists, and possibly creating an improved online musical profile on sites like Last.fm.

Project Results:
  • automatic rating mechanisms
  • automatic playlists
  • implementation of audioscrobbler protocol for submitting to online profile (last.fm)
Required Skills
  • Good programming skills
  • Previous experience with iTunes plugins is a plus

Music Recommendation Database [Archived]

music_rec_db

Title: Music Recommendation Database
Duration: One Semester
Responsible TA: Nicolas Jones
Goals: Creating a music recommender system. Setup a music database  as a first step.
Keywords: data collection, recommender systems
Abstract:

Recommender systems (RS) have started becoming popular to help users overcome the information overload problems on internet. Today’s trends of Web 2.0 and the emergence of social interactions, where content is created and shared by users, have lead to new contexts where RS are used such as music. The goal of the project is create a music database, usable for a future recommender system. The first challenge will be to organise and classify a comprehensive collection of music files (+10000 songs), leading to the second challenge ofcompleting the meta-data of each track, such as artist, track, album, etc. (through automated querying of the Gracenote database, or other) leading to the establishement of a rich and complete music database. The developpment should include the interface for listening and importing new songs. Finally, if time permits, the project will include the setting up of a basic music recommender. Based on the remaining time, the project can be extended such as to for example include data mining goals for add taging values to the songs. Many music websites provide web services for consulting tags related to tracks, and this could be used to create a second level abstraction, an ontology, for enhancing music recommendations.

Project Results:
  • Erection of collection of music files
  • Mapping from a music metadata database
  • Song recognition and assimilation of metadata
  • Importation interface
  • Software architecture for classifying files
Required Skills
  • Data parsing (java, scala or other)
  • Database manipulation (php/mysql)
  • Good Software engeneering skils

E-Commerce Shopping System for Mobile Viewers [Archived]

screenshot-mobile

Title: E-Commerce Shopping System for Mobile Viewers
Duration: One Semester
Responsible TA: Jiyong Zhang
Goals: Design and implement a prototype e-commerce shopping website for mobile users.
Keywords: e-commerce, interaction design, plugin development
Abstract:

Over the past few years there has been an exponential growth both in the number of mobile users. Mobile internet enables users to access online information from mobile devices. An upcoming trend for mobile users would be to buy products directly through their mobiles from anywhere, at anytime.

In this project we intend to build or adapt a shopping website for mobile users. The system can be developed based on some existing open source shopping systems such as osCommerce (http://www.oscommerce.com/). We need to design a module so that the interface can be outputted adequately on mobile devices. This can be done by detecting the connecting device and modifying the stylesheets accordingly, or can also be based on the Wireless Application Protocol (WAP) and creation of wml pages. (The WAP is the world standard for the presentation and delivery of wireless information and telephony services on mobile phones and other wireless terminals, but more and more phones just integrate a light-version web browser). The challenge would be to design interfaces that can be conveniently used by mobile users, and to provide as a plugin/module for an existing and popular shopping system.

Project Results:
  • Implementation of a prototype mobile shopping system;
  • Usability issues of the mobile shopping system;
  • Publication of a plugin
  • Results analysis and report.
Required Skills
  • Previous experience of website development
  • CSS / PHP / Mysql
  • Development experience with CMS systems is welcome (optional)

Implementation of an interactive travel map [Archived]

itmap

Title: Implementation of an interactive travel map
Duration: One Semester
Responsible TA: Li Chen
Goals: Implement a prototype that could visualize the travel-related reviews
Keywords: usability study, visual analytics, visualization
Abstract:

The most distinguishable feature of Web2.0 would be it simulates a platform where users could freely share their reviews, experiences and stories with others. The increasing amount of the sharable information, however, sometimes overwhelms the average users since they could not effectively and efficiently obtain the most useful information to them.

The aim of this project is to implement a prototype that could visualize the travel-related reviews, for example, the users’ recommendations on hotel, sight, shop and restaurant, on a map (e.g. Geneva city map). Therefore, with the map, the tourists can intuitionally get an overview of all the relevant spots with locations and icons representing their corresponding rates. The users could further specify their criteria on the type of spot they are interested in (e.g. filtering the map only leaving the shops that sell Swiss watches and are with average rate higher than 3). They could also move onto a specific recommendation to see its detail (e.g. the restaurant’s contact info).

We believe that this kind of interactive travel map would be much more useful and effective for tourists to search for information, compared to listing all of the textual articles for them to read one by one. Welcome to join the project if you are interested!

Project Results:
  • Collection of basic travel information
  • Prototype implementation
  • Interface usability inspection
Required Skills
  • Previous experience of graphical application development
  • Programming languages: Javascript/ HTML/ PHP
  • Database manipulation

Cultural influences on users’ disposition for decision accuracy and trust in e-commerce [Archived]

Title: Cultural influences on users’ disposition for decision accuracy and trust in e-commerce
Duration: One Semester
Responsible TA: Li Chen
Goals: The goal of this project is to design and implement a prototype vacation package recommendation system through blog articles that the users are interested.
Keywords: qualitative user study design, statistical analysis
Abstract:

In recent years, although many researches have emerged to study the antecedents and subsequences of user trust in online environments, they were mostly targeted to a relatively narrow range of people, i.e. the people from similar cultures or same nationality. Few works have in depth investigated whether the cultural background will have effect on users’ decision behavior and more importantly their subjective perception formation in e-commerce. The goal of the project is therefore to resolve the problem by means of user interviews and qualitative survey. We aim to revealing the actual role of cultural background in users’ disposition for decision accuracy and trust, so as to finally identifying the set of website design guidelines to be useful for a more general range of consumers.

Project Results:
  • Implementation of online questionnaire;
  • Qualitative user study design;
  • Participants recruitment;
  • Results analysis and report.
Required Skills
  • Online survey implementation (HTML/PHP)
  • Statistical analysis basis

A Vacation Package Recommendation System based on Word of Mouth Information [Archived]

Title: A Vacation Package Recommendation System based on Word of Mouth Information
Duration: One Semester
Responsible TA: Jiyong Zhang
Goals: The goal of this project is to design and implement a prototype vacation package recommendation system through blog articles that the users are interested..
Keywords: recommender systems
Abstract:

Over the past few years there has been an exponential growth both in the number of bloggers and blog articles. Reading blog articles now becomes a natural activity for many online users and their taste can be largely determined through the articles they had read and rated. It would be quite beneficial if the recommendation system could gather the user’s preferences directly through her activities in reading/rating blogs and generate a personalized recommendation product list to satisfy her needs.

In this project we intend to develop a vacation package recommendation system which can gather users’ preferences through their reading behaviors and ratings of some travel-related blog articles, and then recommend the desired vacation packages to the end-users. The challenge of this project is that we need to make the recommendation cross through the travel blog article space to the vacation package space.

Project Results:
  • Implementation of a prototype Vacation Package Recommendation website with blog functions;
  • A user study to testify the performance of the recommendation system.
Required Skills
  • Previous experience of website development
  • HTML / CSS / PHP / Mysql / Javascript