Loading...
 
Print

Extending a context-centric Learner Context Model with Private and Professional Interests

 
Diploma Thesis or Magister Thesis or Bachelor Thesis or Master Thesis
Name of Diploma/Magister/Bachelor/Master candidate

Context

The future goal is to have a complex lifelong learner model that can deal with every kind of data that could be used to improve the lifelong learning. Today there is no such model available. The Learning Context Project is able to handle arbitrary context data that is collected by smartphones and sent to the server, where the data is stored. This data is modeled using a context model called Learning Context Model. This context model can be used as a basis for a future lifelong learner model. Right know this model can only deal with extrinsic context data. The next step is to extend the Learning Context Model with intrinsic data like interests. The learner’s interests can be collected from many different sources. There exist two web-services that collect interest data. The papers “Lifelong Learner Modeling in Social Networks” (Marius Klein 2014) and “Lifelong Learner Modeling in Academic Networks” (Darko Dugošija 2012) explain how interests can be extracted from these networks. Combining interests and context data allows the system to work with intrinsic and extrinsic information about the user. This allows it to improve the learning experience and lifelong learning of the user.

Task

The Life Long Learner Model should be able to store very different kinds of data. The goal of this service is to provide a data model that can handle context and interest data. The existing context model of “the Learning-Context Project” is used as a basis for this new model. This base model should be extended to be able to handle data about interests of the learner, but the problem is that interest and context data are very different kind of data. Context data can specify extrinsic information about the learner and his environment. This data is specified by a timestamp and the action at this time. Mostly the actual data is used to adapt the learning methods and content to the actual surrounding of the learner. In contrast interest data do not occur at a certain time and have no specific action or event. Interests are the result of a certain behavior over time. A standardized interest model should be created. Problems like Synonymy or Polysemy can occur when modeling semantic data like interests. So the model should be able to deal with this problems. Another problem is how to connect the learner to the interests and how to rank and weight her interests. In general there are two kinds of interests. First, there are topics the user is interested in over a long period of time, maybe even his entire life. Second, there are topics that are only interesting over a short time period. The user can also have more than one interest at a time. So somehow a weighting for the interests is needed to be able to handle many interests at the same time.
The Lifelong Learner Model should be used by other systems and services. So the data model needs to be accessible from the outside. The different services need to be able to send user related data to the system. The Learning Context Project and Lifelong Learner Modeling in Social and Academic Networks are used as example services that provide context and interest data to our service. Our Service should not only store the data but also process the data and make it available to other services. So the system needs an API that makes it accessible from the outside and some algorithms to process and analyze the data.

  • Basis for a Lifelong Open Learner Model
    • User Profile
    • Context data
    • Interests
    • User can view and edit their data
  • Implement the Model
  • Web-Service
    • API
    • Receive and handle data from other services
    • Semantic Analysis of the data
    • Provide data for other services
  • Information Collection
    • App of the “Context Learning Project”
    • WebTrace and PALM
  • Visualization
    • Context data
    • Interest data
  • Security and Privacy
    • User has control over her data
  • Authentication of users and services
  • Use Methods from “The Learning Context Project”

 

Abstract

 

Download

 

Learning Context(external link)
WebTrace(external link)
PALM(external link)

Supervisor

Dr. Mohamed Amine Chatti
Dipl.-Inform. Hendrik Thüs
Christoph Greven, M.Sc. RWTH


Created by greven. Last Modification: Tuesday, 26. August 2014 11:02:09 by brandt.