The Application Of An Affective Computing Algorithm Based On Temperament Type In E-Learning

With the rapid development of information technology and the popularity ofInternet applications, e-Learning system has gradually been aware of and accept bymore and more people, and bring learning a profound impact. For this, the IEEEestablished the e-Learning Technology System Architecture (LTSA) in2003. Thisstandard established a five-layer structure, and LTSA is mainly about the third layer ofthe standards.Since early LTSA standards less involved with the learners emotional, cognitivepsychological factors, the LTSA is lack of expression of the learners emotional factors,led to the e-Learning system designed unable to meet the individual needs ofdifferent learners. Thus some researchers have started to focus on the affectivecomputing used in e-Learning system."Affective computing" concept was proposedby the MIT laboratory Professor Rosalind Pichard in1997, as a branch of artificialintelligence, its purpose is to make the design of systems and equipment to identify,understand and deal with human emotions. The current affective computingresearch focuses on humans understanding and cognitive technology such asemotional speech recognition, facial expression recognition, body posture andmovement recognition.However, the research work of the existing affective computing-basede-Learning system has less involved the learners psychological research.To solve theabove problem, the main tasks of this paper as follows:Adds the personality database on the basis of LTSA, to make up for the lack ofemotional module of the original system architecture. And forms theELTSA(emotional learning technology system Architecture).Put forward a new algorithm, an affective computing algorithm based ontemperament type. This algorithm is based on the type of temperament theory,generates emotional susceptibility matrix, and extracts five emotional featuresaccording to the emotions that may affect in learning, as well as draws personalizedemotional vector by different learners’ temperament type vector. All these reflect the learners’ emotional state, and as the basis of recommended learning, makinglearning more humane.Models the extended system architecture by Multi-Agent theory and usesUnity3D to confirmatory simulation the system and algorithm.