Aristotle University Of Thessaloniki (AUTH)
Στατιστικές μέθοδοι στον προγραμματισμό έργων πληροφοριακών συστημάτων
Abstract
dc:descriptionThe term Information System (IT) is generic and refers to a system that includes all the functions of an organization. In this particular research area, the key component of an information system is the software used in the organization to store and process data but also the automation of the IT processes. Therefore, the Information System is a product that should cover various specifications and quality requirements related to the needs of its application domain.Τhis thesis mainly focuses in the is the definition of priorities in the different software requirements known as requirements prioritization. The large number of requirements and the different views of the employees participating in the survey that define what are the most important and therefore deserve higher priority, make the problem of decision-making extremely difficult and complex. Those data are collected from questionnaires in which special methods are applied (Cumulative Voting). The questionnaires are completed by the stakeholders in such research containing their views coded into numerical "scores" for the prioritization of various requirements. Although the simple sum of the scores from all participants shows which requirements are the most important, the overall information that exists in the data conceals considerable variability. Τhe variables of all the data sets are limited and correlated. In general the normalized form of these data consists of vectors (in the interval [0,1]) that are sum up to one. Special methods of multivariate analysis required for such data, since the usual regularity assumptions and independence does not apply here. These methods will have to answer to several questions related to the level of alignment between the stakeholders, the conflicting nature of the requirements, with the grouping of views and the comparison of views among the inherit groups of stakeholders. Τhe management of software projects, the appropriate allocation of resources in the various software phases in relation to the investigation of the high complexity of the distribution of effort are interesting problems which have not been studied with multivariate methods yet. In this thesis multivariate statistical methods were applied, that have never been used so far in the literature. For instance the statistical analysis framework known as Compositional Data Analysis (CoDA). CoDA was used in data sets which contained proportions similar to what we get from the priority assignment procedures. This analysis includes a wide variety of techniques that aspire to answer critical questions about the priorities of quality requirements as defined by people who behave and express an opinion according to the position they occupy in an organization and the perspective from which the prioritize the quality of the software. The research effort and the statistical analysis through this thesis tried to focus on the human factor and to explore the differences and similarities in views that ultimately determine the quality of a product.Τhe research conducted as part of this thesis had a goal: to use advanced statistical methods to problems of quality that we usually face in information systems.
Degree
thesis:*- Grantor dc:publisher
- Aristotle University Of Thessaloniki (AUTH)
- Year dc:date
- 2015
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Chatzipetrou, Panagiota
- Χατζηπέτρου, Παναγιώτα
Subjects
dc:subject × 16- Ποιότητα
- Πληροφοριακά συστήματα
- Απαιτήσεις ποιότητας
- Κατανομή προσπάθειας
- Ανθρώπινος παράγοντας
- Quality
- Information systems
- Requirements
- Effort distribution
- Human factor
- Φυσικές Επιστήμες
- Μαθηματικά
- Επιστήμη Ηλεκτρονικών Υπολογιστών και Πληροφορική
- Natural Sciences
- Mathematics
- Computer and Information Sciences
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/35912