Belief Rule-based Methodology Research Articles

A Recognition Model of Driving Risk Based on Belief Rule-Base Methodology

This paper aims to recognize driving risks in individual vehicles online based on a data-driven methodology. Existing advanced driver assistance systems (ADAS) have difficulties in effectively processing multi-source heterogeneous driving data. Furthermore, parameters adopted for evaluating the driving risk are limited in these systems. The approach of data-driven modeling is investigated in this study for utilizing the accumulation of on-road driving data. A recognition model of driving risk based on belief rule-base (BRB) methodology is built, predicting driving safety as a function of driver characteristics, vehicle state and road environment conditions. The BRB model was calibrated and validated using on-road data from 30 drivers. The test results show that the recognition accuracy of our proposed model can reach about 90% in all situations with three levels (none, medium, large) of driving risks. Furthermore, the proposed simplified model, which provides real-time operation, is implemented in a vehicle driving simulator as a reference for future ADAS and belongs to research on artificial intelligence (AI) in the automotive field.

A belief rule based expert system for predicting consumer preference in new product development

In the decision making process of new product development, companies need to understand consumer preference for newly developed products. A recently developed belief rule based (BRB) inference methodology is used to formulate the relationship between consumer preference and product attributes. However, when the number of product attributes is large, the methodology encounters the challenge of dealing with an oversized rule base. To overcome the challenge, the paper incorporates factor analysis into the BRB methodology and develops a BRB expert system for predicting consumer preference of a new product. Firstly, a small number of factors are extracted from product attributes by conducting both exploratory and confirmatory factor analysis. Secondly, a belief rule base is constructed to model the causal relationships between the characteristic factors and consumer preference for products using experts’ knowledge. Furthermore, a BRB expert system is developed for predicting consumer preference in new product development, where the factor values transformed from product attributes are taken as inputs. Relevant rules in the system are activated by the input data, and then the activated rules are aggregated using the evidential reasoning (ER) approach to generate the predicted consumer preference for each product. Finally, the BRB expert system is illustrated using the data collected from 100 consumers of several tea stores through a market survey. The results show that the prototype of the BRB expert system has superior fitting capability on training data and high prediction accuracy on testing data, and it has great potential to be applied to consumer preference prediction in new product development.

Identification of uncertain nonlinear systems: Constructing belief rule-based models

The objective of this paper is to construct reliable belief rule-based (BRB) models for the identification of uncertain nonlinear systems. The BRB methodology is developed from the evidential reasoning (ER) approach and traditional IF–THEN rule based system. It can be used to model complicated nonlinear causal relationships between antecedent attributes and consequents under different types of uncertainty. In a BRB model, various types of information and knowledge with uncertainties can be represented using belief structures, and a belief rule is designed with belief degrees embedded in its possible consequents. In this paper, we first introduce the BRB methodology for modelling uncertain nonlinear systems. Then we present a comparative analysis of three BRB identification models through combining the BRB methodology with nonlinear optimisation techniques. The novel BRB identification models using l∞-norm and minimising mean uncertainties in belief rules (MUBR) show remarkable capabilities of capturing the lower and upper bounds of the interval outputs of uncertain nonlinear systems simultaneously. Trade-off analysis between identification accuracy and interval credibility are briefly discussed. Finally, a numerical study of a simplified car dynamics is conducted to demonstrate the capability and effectiveness of the BRB identification models for the modelling and identification of uncertain nonlinear systems.

Fuzzy Failure Mode and Effects Analysis Using Fuzzy Evidential Reasoning and Belief Rule-Based Methodology

The main objective of this paper is to propose a new risk priority model for prioritizing failures in failure mode and effects analysis (FMEA) on the basis of fuzzy evidential reasoning (FER) and belief rule-based (BRB) methodology. The technique is particularly intended to resolve some of the shortcomings in fuzzy FMEA (i.e., fuzzy rule-based) approaches. In the proposed approach, risk factors like occurrence (O), severity (S), and detection (D), along with their relative importance weights, are described using fuzzy belief structures. The FER approach is used to capture and aggregate the diversified, uncertain assessment information given by the FMEA team members; the BRB methodology is used to model the uncertainty, and nonlinear relationships between risk factors and corresponding risk level; and the inference of the rule-based system is implemented using the weighted average-maximum composition algorithm. The Dempster rule of combination is then used to aggregate all relevant rules for assessing and prioritizing the failure modes that have been identified in FMEA. A case study concerning an ocean going fishing vessel in a marine industry is provided and conducted using the proposed model to illustrate its potential applications and benefits.

Belief rule-based methodology for mapping consumer preferences and setting product targets

Rapid and accurate identification of consumer demands and systematic assessment of product quality are essential to success for new product development, in particular for fast moving consumer goods such as food and drink products. This paper reports an investigation into a belief rule-based (BRB) methodology for quality assessment, target setting and consumer preference prediction in retro-fit design of food and drink products. The BRB methodology can be used to represent the relationships between consumer preferences and product attributes, which are complicated and nonlinear. A BRB system can initially be established using expert knowledge and then optimally trained and validated using data generated from consumer or expert panel assessments or from tests and experiments. The established BRBs can then be used to predict the consumer acceptance of new products or set product target values in retro-fit design. The proposed BRB methodology is applied to the design of a lemonade drink product using real data provided by a sensory product manufacturer in the UK. The results show that the BRB methodology can be used to predict consumer preferences with high accuracy and to set optimal target values for product quality improvement.

Consumer preference prediction by using a hybrid evidential reasoning and belief rule-based methodology

Consumer preference prediction is a key factor to the success of new product development. This paper presents a hybrid evidential reasoning (ER) and belief rule-based (BRB) methodology for consumer preference prediction and a novel application to orange juices. The orange juices are distinguished by their values of sensory attributes, which are grouped for simplicity into different categories such as appearance, aroma, texture, flavour, and aftertaste. The ER approach is used to aggregate consumer preferences for category attributes into an overall preference, and the BRB methodology is used to model the casual relationships between category attributes and their sensory attributes. The casual relationships between the overall preference and the sensory attributes of orange juices are trained and tested using real data and memorized for prediction or new product design. A case study involving 16 orange juices is conducted using the proposed hybrid ER and BRB methodology to demonstrate its novel applications. The results show that the hybrid ER and BRB methodology can fit and predict consumer preferences with high accuracy.