Πανεπιστήμιο Μακεδονίας
Behavioral user modeling for web and cloud end-user development environments
Abstract
dc:descriptionThis thesis addresses the challenges of user modeling in cloud and web based End-User Development (EUD) environments by monitoring users’ behavior and taking into account human characteristics. Initially, an innovative development approach is designed to assist end-users in building database-centric applications following the relational scheme. The proposed approach is called Simple-Talking since all technical terminology is replaced by simple words and concepts more familiar to the end-user. The Simple-Talking approach is integrated into a prototype cloud-based EUD tool and is evaluated via an experimental study. The results reveal high scores of user performance without being affected by the user’s expertise level. Impact on performance seems to be exercised by gender, as well as perceived ease of use and perceived usefulness.Secondly, the RULES model is structured. RULES is a new user model consisting of basic gender properties that tend to affect the end-users’ behavior and performance. RULES model includes the five most important end-user features such as: risk perception, self-efficacy, willingness to learn, etc., that resulted from an extensive literature review of popular Human Computer Interaction theories.To model the user based on the RULES properties, a research is developed to implicitly monitor the end-users behavior. For this, methods to monitor mouse, keyboard and eye movements are applied during the user-system interaction. Mouse and keyboard tracking methodology includes the taxonomy of measurable items like mouse events and keystroke dynamics that can be detected in real time on the client-side and stored in a database. The methodology is experimentally evaluated, and the results of the statistical analysis reveal a number of significant correlations between the mouse/ keyboard metrics and the properties of the RULES model.Similarly, the methodology for recording eye movements includes the classification of eye-metrics such as fixation and pupil diameter. The eye tracking approach is experimentally evaluated on a sample of volunteers using the Tobii Eye Tracker software and hardware tool. Significant correlations between eye behavioral data and properties of the RULES model are detected during the data analysis. According to the thesis findings, gender and behavioral metrics (as derived from mouse/keyboard and eye monitoring research) are key EUD attributes to define end-user developer models.The proposed user model is semantically represented via an ontology-based framework. The model is further exploited to define an inference mechanism by which the collected data can be used in diagnosing users’ domain knowledge level, perceptional and acceptance characteristics. This approach ensures that user modeling is possible when initial knowledge is small and new behaviors are encountered over time.
Degree
thesis:*- Grantor dc:publisher
- Πανεπιστήμιο Μακεδονίας
- Year dc:date
- 2018
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Tzafilkou, Aikaterini
- Τζαφίλκου, Αικατερίνη
Subjects
dc:subject × 15- Μοντελοποίηση χρήστη
- Αλληλεπίδραση ανθρώπου υπολογιστή
- Ανάπτυξη εφαρμογών από τελικούς χρήστες
- Σημασιολογικός ιστός
- Καταγραφή κινήσεων ποντικίου / ματιού
- User modeling
- Human computer interaction
- End - user development
- Semantic web
- Mouse tracking
- Eye tracking
- Φυσικές Επιστήμες
- Επιστήμη Ηλεκτρονικών Υπολογιστών και Πληροφορική
- Natural Sciences
- Computer and Information Sciences
Rights
dc:rights- Statement dc:rights
-
- BY_NC
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/44633