The prediction of traffic accident duration is great significant for rapid disposal of traffic accidents, especially for fast rescue of traffic accidents and re- moving traffic safety hazards. In this paper, two methods, which are based on artificial neural network (ANN) and support vector machine (SVM), are adopted for the accident duration prediction. The proposed method is demonstrated by a case study using data on approximately 235 accidents that occurred on freeways located between Dalian and Shenyang, from 2012 to 2014. The mean absolute error (MAE), the root mean square error (RMSE) and the mean absolute percentage error (MAPE) are used to evaluate the performances of the two measures. The conclusions are as follows: Both ANN and SVM models had the ability to predict traffic accident duration within acceptable limits. The ANN model gets a better result for long duration incident cases. The comprehensive performance of the SVM model is better than the ANN model for the traffic accident duration prediction.
This paper presents development of a day ahead load forecasting (DALF) model for Turkish power system with an artificial neural network (ANN). Effects of special holidays including national and religious days, and hourly random load deviations observed in Turkish power system due to significant arc furnace loads are discussed. Performance of the ANN is investigated in the sense of both DALF performance - in terms of both daily mean absolute percentage error (MAPE) and hourly absolute percentage error (APE) - and hourly secondary reserves required to ensure supply/demand adequacy of the system. The most sensitive cities to DALF in terms of daily city temperature forecasts are ranked in order to reduce the input of the developed ANN and thereby to improve execution of the model. Candidate cities are determined based on both their placement with respect to climatic zones of the country and their contribution to the system load during peak hours. The results show that, although a well-trained ANN could provide very satisfactory daily MAPEs at non-special days, such as ~1%, the hourly absolute percentage errors (APE) could be significant due to large random load disturbances, which necessitate special attention during the day ahead allocation of hourly secondary reserves. By limiting the temperature data set with major cities, the input of ANN reduces significantly while not disturbing the MAPEs. Main contributions of the study are; addressing both benefits of the prioritizing the cities in a power system in the sense of their temperature forecast effects on the DALF performance and assessing the performance of DALF in the sense of necessary amount of secondary reserves in power systems which include significant random load deviations (e.g., large arc furnace loads).
Profitability of Turkish banking sector gained importance after national and international financial crisis happened in the last decade, which revealed the need to make a research on profitability and the factors determining profitability. In recent years, new techniques of soft computing (SC) like genetic algorithms (GAs), fuzzy logic (FL) and especially artificial neural networks (ANNs) have been applied into the financial domain to solve the domain issues because of their successful applications in nonlinear multivariate situations. An adaptive system was needed due to the fact that insufficient use of application software programs for SC and the fact that single software is only applicable for specific model. Furthermore, even though ANNs have been applied to many areas; little attention has been paid to estimation of bank profitability with ANNs. This article is intended to analyze and estimate the profitability of deposit banks in Turkey with an adaptive software model of ANNs which have not been previously applied for this context, comprehensively. The results from the software model, which processes the factors affecting profitability, indicate that all of the variables used have significant impacts in varying proportions on profitability and that obtained estimations achieved the targeted and acceptable performance of success. This software model is expected to provide easiness on estimating bank profitability, since giving successful estimations and not being affected by user differences. Additionally, it is aimed to construct a software model for being used in different fields of study and financial domain.
A literature survey was conducted to appraise the recent applications of artifical intelligence (AI)-based modeling studies in the environmental engineering field. A number of studies on artificial neural networks (ANN), fuzzy logic and adaptive neuro-fuzzy systems (ANFIS) were reviewed and important aspects of these models were highlighted. The results of the extensive literature survey showed that most AI-based prediction models were implemented for the solution of water/wastewater (55.7%) and air pollution (30.8%) related environmental problems compared to solid waste (13.5%) management studies. The present literature review indicated that among the many types of ANNs, the three-layer feed-forward and back-propagation (FFBP) networks were considered as one of the simplest and the most widely used network type. In general, the Levenberg-Marquardt algorithm (LMA) was found as the best-suited training algorithm for several complex and nonlinear real-life problems of environmental engineering. The literature survey showed that for water and wastewater treatment processes, most of AI-based prediction models were introduced to estimate the performance of various biological and chemical treatment processes, and to control effluent pollutant loads and flowrates from a specific system. In air polution related environmental problems, forecasting of ozone (O3) and nitrogen dioxide (NO2) levels, daily and/or hourly particulate matter (PM2.5) and PM10) emissions, and sulfur dioxide (SO2) and carbon monoxide (CO) concentrations were found to be widely modeled. For solid waste management applications, reseachers conducted studies to model weight of waste generation, solid waste composition, and total rate of waste generation.
Objective: The anxiety of Alzheimer's disease (AD) contributes significantly to decreased quality of life, increased morbidity, higher levels of caregiver distress, and the decision to institutionalize a patient. However, the incidence of anxiety in AD patients hasn't been discussed. In this study, artificial neural networks were used to predict the incidence of anxiety inAD patients.
Methods: A large randomized controlled clinical trial was analyzed in this study, which involved AD patients and caregivers from 6 different sites in the United States. The incidence of anxiety in AD patients was predicted by backpropagation artificial neural networks with one and hidden layers. After cross validation, the Predictive Accuracy (PA) of the models was measured to select the best structure of artificial neural networks.
Results: Among all models for predicting the incidence of anxiety in AD patients, the artificial neural network with respectively 6 and 3 neurons in the first and second hidden layers achieved the highest predictive accuracy of 85.56%.
Conclusions: The incidence of anxiety in AD patients can be predicted by an accuracy of over 80%. When used for anxiety prediction, neural networks with two hidden layers perform better than those with one hidden layer. These findings will benefit the prevention and early intervention of anxiety in Alzheimer's patients.
Patients suffering from Parkinson's disease must periodically undergo a series of tests, usually performed at medical facilities, to diagnose the current state of the disease. Parkinson's disease progression assessment is an important set of procedures that supports the clinical diagnosis. A common part of the diagnostic train is analysis of speech signal to identify the disease-specific communication issues. This contribution describes two types of computational models that map speech signal measurements to clinical outputs. Speech signal samples were acquired through measurements from patients suffering from Parkinson's disease. In addition to direct mapping, the developed systems must be able of generalization so that correct clinical scale values can be predicted from future, previously unseen speech signals. Computational methods considered in this paper are artificial neural networks, particularly feedforward networks with several variants of backpropagation learning algorithm, and adaptive network-based fuzzy inference system (ANFIS). In order to speed up the learning process, some of the algorithms were parallelized. Resulting diagnostic system could be implemented in an embedded form to support individual assessment of Parkinson's disease progression from patients' homes.
In recent years the interest of the investors in efficient methods for the forecasting price trend of a share in financial markets has grown steadily. The aim is to accurately forecast the future behavior of the market in order to identificate the so-called "correct timing".
In this paper we analyze three different approaches for forecasting financial data: Autoregression, artificial neural networks and support vector machines and we will determine potentials and limits of these methods. Application to the Italian financial market is also presented.
The process of manual species identification is a daunting task, so much so that the number of taxonomists is seen to be declining. In order to assist taxonomists, many methods and algorithms have been proposed to develop semi-automated and fully automated systems for species identification. While semi-automated tools would require manual intervention by a domain expert, fully automated tools are assumed to be not as reliable as manual or semi-automated identification tools. Hence, in this study we investigate the accuracy of fully automated and semi-automated models for species identification. We have built fully automated and semi-automated species classification models using the monogenean species image dataset. With respect to monogeneans’ morphology, they are differentiated based on the morphological characteristics of haptoral bars, an-chors, marginal hooks and reproductive organs (male and female copulatory organs). Landmarks (in the semi-automated model) and shape morphometric features (in the fully automated model) were extracted from four monogenean species images, which were then classified using k-nearest neighbour and artificial neural network. In semi-automated models, a classification accuracy of 96.67 % was obtained using the k-nearest neighbour and 97.5 % using the artificial neural network, whereas in fully automated models, a classification accuracy of 90 % was obtained using the k-nearest neighbour and 98.8 % using the artificial neural network. As for the cross-validation, semi-automated models performed at 91.2 %, whereas fully automated models performed slightly higher at 93.75 %. and Corresponding author: Sarinder Kaur A/p Kashmir Singh
The detection of insulation failures in buildings could potentially conserve energy supplies and improve future designs. Improvements to thermal insulation in buildings include the development of models to assess fabric gain -heat flux through exterior walls in the building- and heating processes. Thermal insulation standards are now contractual obligations in new buildings, and the energy efficiency of buildings constructed prior to these regulations has yet to be determined. The main assumption is that it will be based on heat flux and conductivity measurement. Diagnostic systems to detect thermal insulation failures should recognize anomalous situations in a building that relate to insulation, heating and ventilation. This highly relevant issue in the construction sector today is approached through a novel intelligent procedure that can be programmed according to local building and heating system regulations and the specific features of a given climate zone. It is based on the following phases. Firstly, the dynamic thermal performance of dif\-ferent variables is specifically modeled. Secondly, an exploratory projection pursuit method called Cooperative Maximum-Likelihood Hebbian Learning extracts the relevant features. Finally, a supervised neural model and identification techniques constitute the model for the diagnosis of thermal insulation failures in building due to the heat flux through exterior walls, using relevant features of the data set. The reliability of the proposed method is validated with real datasets from several Spanish cities in winter time.
This work concentrates on a novel method for empirical estimation
of generalization ability of neural networks. Given a set of training (and testing) data, one can choose a network architecture (nurnber of layers, number of neurons in each layer etc.), an initialization method, and a learning algorithrn to obtain a network. One measure of the performance of a trained network is how dosely its actual output approximates the desired output for an input that it has never seen before. Current methods provide a “number” that indicates the estimation of the generalization ability of the network. However, this number provides no further inforrnation to understand the contributing factors when the generalization ability is not very good. The method proposed uses a number of parameters to define the generalization ability. A set of the values of these parameters provide an estimate of the generalization ability. In addition, the value of each pararneter indicates the contribution of such factors as network architecture, initialization method, training data set, etc. Furthermore, a method has been developed to verify the validity of the estimated values of the parameters.