Skip to main content

Trust based Decision Support Systems for Social Networks with Uncertain Knowledge

Periodic Reporting for period 1 - DeciTrustNET (Trust based Decision Support Systems for Social Networks with Uncertain Knowledge)

Reporting period: 2017-10-04 to 2019-10-03

The world wide success of large scale social information systems such as e-commerce platforms, facilities sharing communities and social networks, make them a very promising paradigm for large scale information sharing and management.
However their anonymity and distributed nature may contribute to the propagation of low quality information, attacks and manipulations.
In this research proposal we have created a novel trust and reputation based framework for social choice in such networks. The proposed system takes into consideration the users’ relationships, the historic evolution of their reputations and their profile similarity to develop a tamper resilient network that guarantees trustworthy communications and transactions.
This model paves the way for an entirely new approach of exploiting the massive information stored in social networks to develop an automated trust based decision-support system under uncertainty and incomplete information. The proposed application has great outreach/commercialisation potential in both e-health environments and marketing and recommender systems.
•Establish a new SNA framework for managing multiple inconsistent heterogeneous information sources that enables the implementation of trust.
•To define trust propagation and aggregation operators for trust networks.
•To create a trust based feedback mechanism to provide personalised recommendations
•To develop a mobile e-health platform based on the proposed trust based social network.
The main research work carried out is composed of three research milestones, RM.
RM1 For this RM, we have developed a new similarity based influence social network that leverage the knowledge of the crowds to model the public opinions dynamic and to reach consensus. The network classifies the agents in different profiles pointing out the influencers and allocating them a preponderant position in the network, and isolating those that may present a malicious behavior, as it is depicted on Fig. 1. A general overview of the proposed system is presented in Fig. 2.
Figure 1 Information spreading scheme
Figure 2 Confidence/Consistency/Similarity network construction approach
Results RM1:
-Model simulation implemented in R
-R. Ureña, Francisco Chiclana, Guy Melançon, Enrique Herrera-Viedma, A social network based approach for consensus achievement in multiperson decision making, Information Fusion, Volume 47, 2019, Pages 72-87.
-Ureña R, Kou G, Wu J, Chiclana F, Herrera‐Viedma E. Dealing with incomplete information in linguistic group decision‐making by means of IntervalType‐2 Fuzzy Sets. Int J Intell Syst. 2019;1‐20.

RM2: In this RM we have developed DeciTrustNET, a trust and reputation based framework for social networks that takes into consideration the users relationships, the historic evolution of their reputations and their profile similarity to develop a tamper resilient network that guarantees trustworthy communications and transactions.
In the figure 3 the main elements used to calculate and propagate both global reputation and trust are depicted, whereas Figure 4 shows the architecture of the proposed framework.

Figure 3 DeciTrustNET main conceptual components
Figure 4 DeciTrustNET Architecture

An extensive experimental analysis has been done confirming that the proposed approach supports robust trust and reputation establishment among the users, even in social network under the presence of malicious users. Moreover DeciTrustNET outperforms the state of the art approaches with precision values above 0.75 even when there is an 80% of malicious users.
Results RM2:
-Mathematical model of DeciTrustNET, simulated in R.
-R. Ureña, F. Chiclana, E. Herrera-Viedma DeciTrustNET, graph based trust and reputation framework for
social networks, Information Fusion, 2020, InPress
-R. Ureña, Gang Kou, Yucheng Dong, Francisco Chiclana, Enrique Herrera-Viedma, A review on trust propagation and opinion dynamics in social networks and group decision making frameworks, Information Sciences, Volume 478, 2019, Pages 461-475.

RM3 e-health Android mobile app to monitor both food ingestion and physical activity;

An android mobile application that monitors the food ingestion and the physical activity thanks to a fitbit bracelet has been fully implemented. The architecture of the proposed app is depicted in Figure 5.
Figure 5 E-health mobile app architecture

To better characterised the profile of the aging populations, to provide better recommendations with DecitrustNET we have implemented m-SFT (mobile Senior Fitness Test) a novel m-health app.
Results RM3:
-Android mobile monitoring application food ingestion and exercise (prototype version)
-m-sft: Mobile Senior fitness test (Android m-health system, prototype version)
-Ureña, R.; Chiclana, F.; Gonzalez-Alvarez, A.; Herrera-Viedma, E.; Moral-Munoz, J.A. m-SFT: A Novel Mobile Health System to Assess the Elderly Physical Condition. Sensors 2020, 20, 1462.
-R. Ureña, A. Gonzalez-Alvarez, F. Chiclana, E. Herrera-Viedma, J.A. Moral-Munoz: Intelligent m-Health App to Evaluate the Elderly Physical Condition. SoMeT 2018, Granada, Spain, 87-100

Dissemination & Exploitation

Website :
ResearchGate URL
Academic publications
The results from this project have been published in high impact scientific journals and the articles have obtained great attention from scientific community, the articles published in 2019 have already been cited more than 100 times in Google Scholar.
In order to allow open access all publications derived from the project have been deposited in the De Montfort Open Research Archive ( online research repository.

Network Meetings: Marie Curie Alumni Association annual conference, UK-Marie Curie Alumni Association annual event, ESOF conference 2018 and the ICT-Europe-2018.

Special session organisation
In collaboration with well known researchers in the area of Intelligent decision making in social networks two special sessions have been organised:
-Social Networks and Collaborative Decision Making, (ITQM 2018), Nebraska, US, October 2018.
-Intelligent decision making and consensus in social networks, IEEE System Man and Cybernetics IEEE SMC 2019 October Bari, Italy.
Given Seminars:
Dr Urena has been invited to to present DeciTrustNET in a number of seminars in the UK, in EON Germany and in the Institut national de la santé et de la recherche médicale, Inserm, France.
DeciTrustNET allows robust trust and reputation based communication between agents in a network even in the presence of malicious and new users. This new framework proposes an entirely new form of exploiting the large amount of information stored in social network to carry out automated decision-support under uncertainty and incomplete information. The propose framework is particularly useful in three specific scenarios: Consensus achievement in group decision making processes such as the ones carried out in e-democracy; e-health platforms to provide recommendations on how to keep a healthy lifestyle; and in recommender systems for e-commerce and e-marketing.
An extension of the project DeciTrustNET and its application to Cancer recovering patients in a large scale is being considered in the context of a research collaboration with Inserm researchers.