Narrow Results

In some practical situations – e.g., when treating a new illness – we do not have enough data to make valid statistical conclusions. In such situations, it is necessary to use expert knowledge – and thus, it is beneficial to use fuzzy techniques that were specifically designed to process such knowledge. At first glance, it may seem that in situations when we have large amounts of data, the relative importance of expert knowledge should decrease. However, somewhat surprisingly, it turns out that expert knowledge is still very useful in the current age of big data. In this paper, we explain how exactly (and why) expert knowledge is useful, and we overview efficient methods for processing this knowledge. This overview is illustrated by examples from environmental science, geosciences, engineering (in particular, aircraft maintenance and underwater robots), and medicine.

Expert knowledge has been proved by substantial studies to be contributory to higher forecasting performance; meanwhile, its application is criticized and opposed by some groups for biases and inconsistency inherent in experts’ subjective judgment. This paper proposes a new approach to improving forecasting performance, which takes advantage of expert knowledge by constructing a constraint equation rather than directly adjusting the predicted values by experts. For the comparison purpose, the proposed approach, together with several widely used models including ARIMA, BP-ANN and the judgment model (JM), is applied to forecasting the container throughput of Guangzhou Port, which is one of the most important ports of China. Forecasting performances of the above models are compared and the results clearly show superiority of the proposed approach over its rivals, which implies that expert knowledge will make positive contribution as long as it is used in a right way.

One of the main challenges in the Multi-Criteria Decision Analysis (MCDA) field is how we can identify criteria weights correctly. However, some MCDA methods do not use an explicitly defined vector of criterion weights, leaving the decision-maker lacking knowledge in this area. This is the motivation for our research because, in that case, a decision-maker cannot indicate a detailed justification for the proposed results. In this paper, we focus on the problem of identifying criterion weights in multi-criteria problems. Based on the proposed Characteristic Object Method (COMET) model, we used linear regression to determine the global and local criterion weights in the given situation. The work was directed toward a practical problem, i.e., evaluating Formula One drivers’ performances in races in the 2021 season. The use of the linear regression model allowed for identifying the criterion weights. Thanks to that, the expert using the system based on the COMET method can be equipped with the missing knowledge about the significance of the criteria. The local identification allowed us to establish how small input parameter changes affect the final result. However, the local weights are still highly correlated with global weights. The proposed approach to identifying weights proved to be an effective tool that can be used to fill in the missing knowledge that the expert can use to justify the results in detail. Moreover, weights identified in that way seem to be more reliable than in the classical approach, where we know only global weights. From the research it can be concluded, that the identified global and local weights importance provide highly similar results, while the former one provides more detailed information for the expert. Furthermore, the proposed approach can be used as a support tool in the practical problem as it guarantees additional data for the decision-maker.

Many traditional pruning methods assume that all the datasets are equally probable and equally important, so they apply equal pruning to all the datasets. However, in real-world classification problems, all the datasets are not equal and considering equal pruning rate during pruning tends to generate a decision tree with a large size and high misclassification rate.

In this paper, we present a practical algorithm to deal with the data specific classification problem when there are datasets with different properties. Another key motivation of the data specific pruning in the paper is "trading accuracy and size". A new algorithm called Expert Knowledge Based Pruning (EKBP) is proposed to solve this dilemma. We proposed to integrate error rate, missing values and expert judgment as factors for determining data specific pruning for each dataset. We show by analysis and experiments that using this pruning, we can scale both accuracy and generalisation for the tree that is generated. Moreover, the method can be very effective for high dimensional datasets. We conduct an extensive experimental study on openly available 40 real world datasets from UCI repository. In all these experiments, the proposed approach shows considerably reduction of tree size having equal or better accuracy compared to several benchmark decision tree methods that are proposed in literature.

Most existing automatic train operation (ATO) models are based on different train control algorithms and aim to closely track the target velocity curve optimized offline. This kind of model easily leads to some problems, such as frequent changes of the control outputs, inflexibility of real-time adjustments, reduced riding comfort and increased energy consumption. A new data-driven train operation (DTO) model is proposed in this paper to conduct the train control by employing expert knowledge learned from experienced drivers, online optimization approach based on gradient descent, and a heuristic parking method. Rather than directly to model the target velocity curve, the DTO model alternatively uses the online and offline operation data to infer the basic control output according to the domain expert knowledge. The online adjustment is performed over the basic output to achieve stability. The proposed train operation model is evaluated in a simulation platform using the field data collected in YiZhuang Line of Beijing Subway. Compared with the curve tracking approaches, the proposed DTO model achieves significant improvements in energy consumption and riding comfort. Furthermore, the DTO model has more advantages including the flexibility of the timetable adjustments and the less operation mode conversions, that are beneficial to the service life of train operation systems. The DTO model also shows velocity trajectories and operation mode conversions similar to the one of experienced drivers, while achieving less energy consumption and smaller parking error. The robustness of the proposed algorithm is verified through numerical simulations with different system parameters, complicated velocity restrictions, diverse running times and steep gradients.

This paper addresses the problem of car headlight lens inspection. First, the currently quality control of lenses with the defect characterization is presented. Second, a vision-sensor-planning system is developed. This system utilizes the CAD information of the headlight lens and the camera model to plan camera viewpoints. The desirable sensor poses are achieved by a genetic algorithm. To improve the performance of the system the customer requirements and the skill of the human inspector are included through a fuzzy system.

In this paper auxiliary information on laboratories is combined with proficiency testing (PT) data to compute more reliable consensus values and associated uncertainties. This new methodology extensively relies on expert knowledge to assume measurement bias, to investigate the sources of measurement bias, to model relationships between sources of bias and to make hypotheses on the form of the consensus value computed as a weighted mean of the measurement results. This approach thus extends the modelling of PT data to the modelling of measurement bias. As such, the approach provides additional results aimed at diagnosing major sources of measurement bias. The full methodology is applied to a PT involving environmental laboratories measuring water pollutants.