Fuzzy Cognitive Map-Driven Comprehensive Time-Series Classification.

This article presents a comprehensive approach for time-series classification. The proposed model employs a fuzzy cognitive map (FCM) as a classification engine. Preprocessed input data feed the employed FCM. Map responses, after a postprocessing procedure, are used in the calculation of the final classification decision. The time-series data are staged using the moving-window technique to capture the time flow in the training procedure. We use a backward error propagation algorithm to compute the required model hyperparameters. Four model hyperparameters require tuning. Two are crucial for the model construction: 1) FCM size (number of concepts) and 2) window size (for the moving-window technique). Other two are important for training the model: 1) the number of epochs and 2) the learning rate (for training). Two distinguishing aspects of the proposed model are worth noting: 1) the separation of the classification engine from pre- and post-processing and 2) the time flow capture for data from concept space. The proposed classifier joins the key advantage of the FCM model, which is the interpretability of the model, with the superior classification performance attributed to the specially designed pre- and postprocessing stages. This article presents the experiments performed, demonstrating that the proposed model performs well against a wide range of state-of-the-art time-series classification algorithms.

Construction and Supervised Learning of Long-Term Grey Cognitive Networks.

Modeling a real-world system by means of a neural model involves numerous challenges that range from formulating transparent knowledge representations to obtaining reliable simulation errors. However, that knowledge is often difficult to formalize in a precise way using crisp numbers. In this paper, we present the long-term grey cognitive networks which expands the recently proposed long-term cognitive networks (LTCNs) with grey numbers. One advantage of our neural system is that it allows embedding knowledge into the network using weights and constricted neurons. In addition, we propose two procedures to construct the network in situations where only historical data are available, and a regularization method that is coupled with a nonsynaptic backpropagation algorithm. The results have shown that our proposal outperforms the LTCN model and other state-of-the-art methods in terms of accuracy.

Nonsynaptic Error Backpropagation in Long-Term Cognitive Networks.

We introduce a neural cognitive mapping technique named long-term cognitive network (LTCN) that is able to memorize long-term dependencies between a sequence of input and output vectors, especially in those scenarios that require predicting the values of multiple dependent variables at the same time. The proposed technique is an extension of a recently proposed method named short-term cognitive network that aims at preserving the expert knowledge encoded in the weight matrix while optimizing the nonlinear mappings provided by the transfer function of each neuron. A nonsynaptic, backpropagation-based learning algorithm powered by stochastic gradient descent is put forward to iteratively optimize four parameters of the generalized sigmoid transfer function associated with each neuron. Numerical simulations over 35 multivariate regression and pattern completion data sets confirm that the proposed LTCN algorithm attains statistically significant performance differences with respect to other well-known state-of-the-art methods.

Deterministic learning of hybrid Fuzzy Cognitive Maps and network reduction approaches.

Hybrid artificial intelligence deals with the construction of intelligent systems by relying on both human knowledge and historical data records. In this paper, we approach this problem from a neural perspective, particularly when modeling and simulating dynamic systems. Firstly, we propose a Fuzzy Cognitive Map architecture in which experts are requested to define the interaction among the input neurons. As a second contribution, we introduce a fast and deterministic learning rule to compute the weights among input and output neurons. This parameterless learning method is based on the Moore-Penrose inverse and it can be performed in a single step. In addition, we discuss a model to determine the relevance of weights, which allows us to better understand the system. Last but not least, we introduce two calibration methods to adjust the model after the removal of potentially superfluous weights.

Short-term cognitive networks, flexible reasoning and nonsynaptic learning.

While the machine learning literature dedicated to fully automated reasoning algorithms is abundant, the number of methods enabling the inference process on the basis of previously defined knowledge structures is scanter. Fuzzy Cognitive Maps (FCMs) are recurrent neural networks that can be exploited towards this goal because of their flexibility to handle external knowledge. However, FCMs suffer from a number of issues that range from the limited prediction horizon to the absence of theoretically sound learning algorithms able to produce accurate predictions. In this paper we propose a neural system named Short-term Cognitive Networks that tackle some of these limitations. In our model, used for regression and pattern completion, weights are not constricted and may have a causal nature or not. As a second contribution, we present a nonsynaptic learning algorithm to improve the network performance without modifying the previously defined weight matrix. Besides, we derive a stop condition to prevent the algorithm from iterating without significantly decreasing the global simulation error.

Traffic accident segmentation by means of latent class clustering.

Traffic accident data are often heterogeneous, which can cause certain relationships to remain hidden. Therefore, traffic accident analysis is often performed on a small subset of traffic accidents or several models are built for various traffic accident types. In this paper, we examine the effectiveness of a clustering technique, i.e. latent class clustering, for identifying homogenous traffic accident types. Firstly, a heterogeneous traffic accident data set is segmented into seven clusters, which are translated into seven traffic accident types. Secondly, injury analysis is performed for each cluster. The results of these cluster-based analyses are compared with the results of a full-data analysis. This shows that applying latent class clustering as a preliminary analysis can reveal hidden relationships and can help the domain expert or traffic safety researcher to segment traffic accidents.

Fuzzy-Rough Cognitive Networks.

Rough Cognitive Networks (RCNs) are a kind of granular neural network that augments the reasoning rule present in Fuzzy Cognitive Maps with crisp information granules coming from Rough Set Theory. While RCNs have shown promise in solving different classification problems, this model is still very sensitive to the similarity threshold upon which the rough information granules are built. In this paper, we cast the RCN model within the framework of fuzzy rough sets in an attempt to eliminate the need for a user-specified similarity threshold while retaining the model's discriminatory power. As far as we know, this is the first study that brings fuzzy sets into the domain of rough cognitive mapping. Numerical results in the presence of 140 well-known pattern classification problems reveal that our approach, referred to as Fuzzy-Rough Cognitive Networks, is capable of outperforming most traditional classifiers used for benchmarking purposes. Furthermore, we explore the impact of using different heterogeneous distance functions and fuzzy operators over the performance of our granular neural network.

Ranking and selecting dangerous crash locations: correcting for the number of passengers and Bayesian ranking plots.

INTRODUCTION: In this paper a sensitivity analysis is performed to investigate how big the impact would be on the current ranking of crash locations in Flanders (Belgium) when only taking into account the most serious injury per crash instead of all the injured occupants. RESULTS: Results show that this would lead to a different selection of 23.8% of the 800 sites that are currently considered as dangerous. CONCLUSIONS: Considering this impact quantity, the researchers want to sensitize government that giving weight to the severity of the crash can correct for the bias that occurs when the number of occupants of the vehicles are subject to coincidence. Additionally, probability plots are generated to provide policy makers with a scientific instrument with intuitive appeal to select dangerous road locations on a statistically sound basis. Impact on industry Considering the impact quantity of giving weight to the severity of the crash instead of to all the injured occupants of the vehicle on the ranking of crash sites, the authors want to sensitize government to carefully choose the criteria for ranking and selecting crash locations in order to achieve an enduring and successful traffic safety policy. Indeed, giving weight to the severity of the crash can correct for the bias that occurs when the number of occupants of the vehicles are subject to coincidence. However, it is up to the government to decide which priorities should be stressed in the traffic safety policy. Then, the appropriate weighting value combination can be chosen to rank and select the most dangerous crash locations. Additionally, the probability plots proposed in this paper can provide policy makers with a scientific instrument with intuitive appeal to select dangerous road locations on a statistically sound basis. Note that, in practice, one should not only rank the crash locations based on the benefits that can be achieved from tackling these locations. Future research is also needed to incorporate the costs of infrastructure measures and other actions that these crash sites require in order to enhance the safety on these locations. By balancing these costs and benefits against each other, the crash locations can then be ranked according to the order in which they should be prioritized.

Road safety risk evaluation and target setting using data envelopment analysis and its extensions.

Currently, comparison between countries in terms of their road safety performance is widely conducted in order to better understand one's own safety situation and to learn from those best-performing countries by indicating practical targets and formulating action programmes. In this respect, crash data such as the number of road fatalities and casualties are mostly investigated. However, the absolute numbers are not directly comparable between countries. Therefore, the concept of risk, which is defined as the ratio of road safety outcomes and some measure of exposure (e.g., the population size, the number of registered vehicles, or distance travelled), is often used in the context of benchmarking. Nevertheless, these risk indicators are not consistent in most cases. In other words, countries may have different evaluation results or ranking positions using different exposure information. In this study, data envelopment analysis (DEA) as a performance measurement technique is investigated to provide an overall perspective on a country's road safety situation, and further assess whether the road safety outcomes registered in a country correspond to the numbers that can be expected based on the level of exposure. In doing so, three model extensions are considered, which are the DEA based road safety model (DEA-RS), the cross-efficiency method, and the categorical DEA model. Using the measures of exposure to risk as the model's input and the number of road fatalities as output, an overall road safety efficiency score is computed for the 27 European Union (EU) countries based on the DEA-RS model, and the ranking of countries in accordance with their cross-efficiency scores is evaluated. Furthermore, after applying clustering analysis to group countries with inherent similarity in their practices, the categorical DEA-RS model is adopted to identify best-performing and underperforming countries in each cluster, as well as the reference sets or benchmarks for those underperforming ones. More importantly, the extent to which each reference set could be learned from is specified, and practical yet challenging targets are given for each underperforming country, which enables policymakers to recognize the gap with those best-performing countries and further develop their own road safety policy.

Benchmarking road safety: lessons to learn from a data envelopment analysis.

Road safety performance indicators (SPI) have recently been proposed as a useful instrument in comparing countries on the performance of different risk aspects of their road safety system. In this respect, SPIs should be actionable, i.e. they should provide clear directions for policymakers about what action is needed and which priorities should be set in order to improve a country's road safety level in the most efficient way. This paper aims at contributing to this issue by proposing a computational model based on data envelopment analysis (DEA). Based on the model output, the good and bad aspects of road safety are identified for each country. Moreover, targets and priorities for policy actions can be set. As our data set contains 21 European countries for which a separate, best possible model is constructed, a number of country-specific policy actions can be recommended. Conclusions are drawn regarding the following performance indicators: alcohol and drugs, speed, protective systems, vehicle, infrastructure and trauma management. For each country that performs relatively poor, a particular country will be assigned as a useful benchmark.