IADIS-INTERNATIONAL JOURNAL ON COMPUTER SCIENCE AND INFORMATION SYSTEMS

12

Authors: Marcelo Henriques de Brito and Paula Esteban do Valle Jardim

This work presents a new approach to behavioral finance with a theoretical contribution by suggesting and discussing with examples a list of group behavioral biases along with established individual behavioral biases, bringing, hence, an additional outlook on how behavioral biases affect financial decisions. While individual behavioral biases are detected in individuals acting alone, group behavioral biases require the scrutiny of group behavior. This awareness may be particularly important to institutional investors, whose decisions basically stem from a committee or a group that will exhibit behavioral biases depending on how the group members interact between themselves when making a decision, which may include negotiation activities and not necessarily be related to personality or hierarchy. The focus on individual investors deciding on personal investments explain the need of work already developed in behavioral finance, which focus on individual behavioral biases, which may be a consequence from either cognitive errors or emotional biases. However, decisions from institutional investors basically stem from a committee or a group that will exhibit behavioral biases depending on how the group members interact between themselves when making a decision. To address the challenge of identifying causes and consequences for unexpected or unsuitable financial decision-making within a group, this work initially retrieves previous work on individual behavioral biases, linking emotional biases and cognitive errors to the “system 1” and “system 2” decision-making framework. Then, a conceptual contribution of this paper, which may be particularly relevant for institutional investors, is to explain with examples - after research and experience - which are the group behavioral biases and their impact upon financial decisions. Individual behavioral biases already acknowledged in other works on behavioral finance are contrasted in this work to the suggested group behavioral biases. Furthermore, this work suggests that there are two broad types of group behavioral biases: group dynamics biases and information-acceptance biases. Each broad type is subdivided into biases related to the structure of the group and biases related to how the group decision-making procedure occurs. Group dynamics biases related to the manner the group is structured are the following: kin bias (belonging bias), harmony bias, and competition bias. On the other hand, group dynamics biases may be sorted according to five different decision-making procedures, namely: herding, fad bias, Plato bias (denial bias), scarcity bias, and home bias. Information-acceptance biases may also be divided into a set related to how the group is structured (comprising: gender bias, survivorship bias, and equality bias) and another set related to the decision-making procedure (comprising cohesion bias, analytical bias, momentum bias, halo effect bias, and deceit bias). All these group bias share errors related to collecting, processing, or disclosing information. Normally the group members are conscious on these biases but are either unable or unwilling to change the setting. This work may trigger ideas for future research and surveys on how groups analyze, recommend, and act on financial information and investment opportunities.

14

Authors: Ruel Welch, Temitope Alade and Lynn Nichol

It has been observed that mobile learning (mLearning) in institutions like Museums in the United Kingdom (UK) has been underutilized. mLearning usage could potentially increase productivity by delivering just-in-time technical knowledge to the science museum group (SMG) staff. This study uses the unified theory of acceptance and use of technology (UTAUT) model to determine factors affecting mLearning adoption at the SMG. Two research questions were formulated based on an adaptation of the UTAUT model. 1) What are the determinants of behavior intentions to use mLearning at the SMG? 2) Does gender or age have a moderating effect on the factors that determine behavior intentions to use mLearning at the SMG? 118 respondents were surveyed from the SMG. Data obtained were analyzed using Structured Equation Modelling on IBM SPSS 20 and Amos version 25. Results indicate that the UTAUT constructs, performance expectancy, effort expectancy, social influence and facilitating conditions are all significant determinants of behavioral intention to use mLearning. A newly proposed construct, self-directed learning was not a significant determinant of behaviour intentions. Further examination found age and gender moderate the relationship between the UTAUT constructs. These findings present several useful implications for mLearning research and practice for ICT service desk at the SMG. The research contributes to mLearning technology adoption and strategy.

15

Authors: Youssouf Ismail Cherifi and Abdelhakim Dahimene

Identifying the speaker has become more of an imperative thing to do in the modern age. Especially since most personal and professional appliances rely on voice commands or speech in general terms to operate. These systems need to discern the identity of the speaker rather than just the words that have been said to be both smart and safe. Especially if we consider the numerous advanced methods that have been developed to generate fake speech segments. The objective of this paper is to improve upon the existing voice-based biometrics to keep up with these synthesizers. The proposed method focuses on defining a novel and more speaker adapted features by implying artificial neural networks and transfer learning. The approach uses pre-trained networks to define a mapping from two complementary acoustic features to a speaker adapted phonetic features. The complementary acoustics features are paired to provide both information about how the speech segments are perceived (type 1 feature) and produced (type 2 feature). The approach was evaluated using both a small and large closed-speaker data set. Primary results are encouraging and confirm the usefulness of such an approach to extract speaker adapted features whether for classical machine learning algorithms or advanced neural structures such as LSTM or CNN.

18

Authors: Alexandre V. Maschio and Nuno M. R. Correia

This article reports a case study in which a digital learning object (DLO) was developed to assist in pedagogical practice in higher education (in the audiovisual area) and presents the technical and theoretical stages of the tool development process (DLO) and its assessment. The objective of the research was mainly to evaluate the pedagogical contribution of DLO through the perception of students who were subdivided into four groups, performed two practical exercises at different times and order, during a 60-hour course. Both exercises of the same complexity were performed without and with the aid of the digital tool. Subsequently, the participants answered forms to be able to evaluate the tool, in addition to having their audiovisual products developed during the course/research been used for a blind analysis to infer qualitative gain at work due to possible time savings generated by the automation function of the tool. In the end, it was found that the DLO tool was very well evaluated conceptually and considered relevant, differentiated, with credibility and high intention of use (among other metrics). The blind analysis showed that there was no qualitative difference due to the possible gain in time between the works developed with or without the aid of the tool.