Search results

Credit scoring is important for financial institutions in order to accurately predict the likelihood of business failure. Related studies have shown that machine learning techniques, such as neural networks, outperform many statistical approaches to solving this type of problem, and advanced machine learning techniques, such as classifier ensembles and hybrid classifiers, provide better prediction performance than single machine learning based classification techniques. However, it is not known which type of advanced classification technique performs better in terms of financial distress prediction. The paper aims to discuss these issues.

This paper compares neural network ensembles and hybrid neural networks over three benchmarking credit scoring related data sets, which are Australian, German, and Japanese data sets.

The experimental results show that hybrid neural networks and neural network ensembles outperform the single neural network. Although hybrid neural networks perform slightly better than neural network ensembles in terms of predication accuracy and errors with two of the data sets, there is no significant difference between the two types of prediction models.

The originality of this paper is in comparing two types of advanced classification techniques, i.e. hybrid and ensemble learning techniques, in terms of financial distress prediction.

Phishing is one of the major threats affecting businesses worldwide in current times. Organizations and customers face the hazards arising out of phishing attacks because of anonymous access to vulnerable details. Such attacks often result in substantial financial losses. Thus, there is a need for effective intrusion detection techniques to identify and possibly nullify the effects of phishing. Classifying phishing and non-phishing web content is a critical task in information security protocols, and full-proof mechanisms have yet to be implemented in practice. The purpose of the current study is to present an ensemble machine learning model for classifying phishing websites.

A publicly available data set comprising 10,068 instances of phishing and legitimate websites was used to build the classifier model. Feature extraction was performed by deploying a group of methods, and relevant features extracted were used for building the model. A twofold ensemble learner was developed by integrating results from random forest (RF) classifier, fed into a feedforward neural network (NN). Performance of the ensemble classifier was validated using k-fold cross-validation. The twofold ensemble learner was implemented as a user-friendly, interactive decision support system for classifying websites as phishing or legitimate ones.

Experimental simulations were performed to access and compare the performance of the ensemble classifiers. The statistical tests estimated that RF_NN model gave superior performance with an accuracy of 93.41 per cent and minimal mean squared error of 0.000026.

The research data set used in this study is publically available and easy to analyze. Comparative analysis with other real-time data sets of recent origin must be performed to ensure generalization of the model against various security breaches. Different variants of phishing threats must be detected rather than focusing particularly toward phishing website detection.

The twofold ensemble model is not applied for classification of phishing websites in any previous studies as per the knowledge of authors.

Classifying the types of fabric defects in the textile industry requires a way to effectively detect. The traditional textile fabric defects detection method is human eyes, which performs very low efficiency and high cost. Therefore, how to improve the classification accuracy of textile fabric defects by using current artificial intelligence and to better meet the needs in the textile industry, the purpose of this article is to develop a method to improve the accuracy of textile fabric defects classification.

To improve the accuracy of textile fabric defects classification, an ensemble learning-based convolutional neural network (CNN) method in terms of textile fabric defects classification (short for ECTFDC) on an enhanced TILDA database is used. ECTFDC first adopts ensemble learning-based model to classify five types of fabric defects from TILDA. Subsequently, ECTFDC extracts features of fabric defects via an ensemble multiple convolutional neural network model and obtains parameters by using transfer learning method.

The authors applied ECTFDC on an enhanced TILDA database to improve the robustness and generalization ability of the proposed networks. Experimental results show that ECTFDC outperforms the other networks, the precision and recall rates are 97.8%, 97.68%, respectively.

The ensemble convolutional neural network textile fabric defect classification method in this paper can quickly and effectively classify textile fabric defect categories; it can reduce the production cost of textiles and it can alleviate the visual fatigue of inspectors working for a long time.

The purpose of this paper is to present a deep ensemble neural network model for the detection of forest fires in aerial vehicle videos.

Presented deep ensemble models include four convolutional neural networks (CNNs): a faster region-based CNN (Faster R-CNN), a simple one-stage object detector (RetinaNet) and two different versions of the you only look once (Yolo) models. The presented method generates its output by fusing the outputs of these different deep learning (DL) models.

The presented fusing approach significantly improves the detection accuracy of fire incidents in the input data.

The computational complexity of the proposed method which is based on combining four different DL models is relatively higher than that of using each of these models individually. On the other hand, however, the performance of the proposed approach is considerably higher than that of any of the four DL models.

The simulation results show that using an ensemble model is quite useful for the precise detection of forest fires in real time through aerial vehicle videos or images.

By this method, forest fires can be detected more efficiently and precisely. Because forests are crucial breathing resources of the earth and a shelter for many living creatures, the social impact of the method can be considered to be very high.

This study fuses the outputs of different DL models into an ensemble model. Hence, the ensemble model provides more potent and beneficial results than any of the single models.

The purpose of this paper is three-fold: to review the categories explaining mainly optimization algorithms (techniques) in that needed to improve the generalization performance and learning speed of the Feedforward Neural Network (FNN); to discover the change in research trends by analyzing all six categories (i.e. gradient learning algorithms for network training, gradient free learning algorithms, optimization algorithms for learning rate, bias and variance (underfitting and overfitting) minimization algorithms, constructive topology neural networks, metaheuristic search algorithms) collectively; and recommend new research directions for researchers and facilitate users to understand algorithms real-world applications in solving complex management, engineering and health sciences problems.

The FNN has gained much attention from researchers to make a more informed decision in the last few decades. The literature survey is focused on the learning algorithms and the optimization techniques proposed in the last three decades. This paper (Part II) is an extension of Part I. For the sake of simplicity, the paper entitled “Machine learning facilitated business intelligence (Part I): Neural networks learning algorithms and applications” is referred to as Part I. To make the study consistent with Part I, the approach and survey methodology in this paper are kept similar to those in Part I.

Combining the work performed in Part I, the authors studied a total of 80 articles through popular keywords searching. The FNN learning algorithms and optimization techniques identified in the selected literature are classified into six categories based on their problem identification, mathematical model, technical reasoning and proposed solution. Previously, in Part I, the two categories focusing on the learning algorithms (i.e. gradient learning algorithms for network training, gradient free learning algorithms) are reviewed with their real-world applications in management, engineering, and health sciences. Therefore, in the current paper, Part II, the remaining four categories, exploring optimization techniques (i.e. optimization algorithms for learning rate, bias and variance (underfitting and overfitting) minimization algorithms, constructive topology neural networks, metaheuristic search algorithms) are studied in detail. The algorithm explanation is made enriched by discussing their technical merits, limitations, and applications in their respective categories. Finally, the authors recommend future new research directions which can contribute to strengthening the literature.

The FNN contributions are rapidly increasing because of its ability to make reliably informed decisions. Like learning algorithms, reviewed in Part I, the focus is to enrich the comprehensive study by reviewing remaining categories focusing on the optimization techniques. However, future efforts may be needed to incorporate other algorithms into identified six categories or suggest new category to continuously monitor the shift in the research trends.

The authors studied the shift in research trend for three decades by collectively analyzing the learning algorithms and optimization techniques with their applications. This may help researchers to identify future research gaps to improve the generalization performance and learning speed, and user to understand the applications areas of the FNN. For instance, research contribution in FNN in the last three decades has changed from complex gradient-based algorithms to gradient free algorithms, trial and error hidden units fixed topology approach to cascade topology, hyperparameters initial guess to analytically calculation and converging algorithms at a global minimum rather than the local minimum.

The existing literature surveys include comparative study of the algorithms, identifying algorithms application areas and focusing on specific techniques in that it may not be able to identify algorithms categories, a shift in research trends over time, application area frequently analyzed, common research gaps and collective future directions. Part I and II attempts to overcome the existing literature surveys limitations by classifying articles into six categories covering a wide range of algorithm proposed to improve the FNN generalization performance and convergence rate. The classification of algorithms into six categories helps to analyze the shift in research trend which makes the classification scheme significant and innovative.

Artificial neural network (ANN) has been used for risk analysis in various applications such as engineering, financial and facilities management. However, use of a single network has become less accurate when the problem is complex with a large number of variables to be considered. Ensemble neural network (ENN) architecture has proposed to overcome these difficulties of solving a complex problem. ENN consists of many small “expert networks” that learn small parts of the complex problem, which are established by decomposing it into its sub levels. This paper seeks to address these issues.

ENN model was developed to analyze risks in maintainability of buildings which is known as a complex problem with a large number of risk variables. The model comprised four expert networks to represent building components of roof, façade, internal areas and basement. The accuracy of the model was tested using two error terms such as network error and generalization error.

The results showed that ENN performed well in solving complex problems by decomposing the problem into its sub levels.

The application of ensemble network would create a new concept of analyzing complex risk analysis problems. The study also provides a useful tool for designers, clients, facilities managers/maintenance managers and users to analyze maintainability risks of buildings at early stages.

The crowdfunding market has experienced rapid growth in recent years. However, not all projects are successfully financed because of information asymmetries between the founder and the providers of external finance. This shortfall in funding has made factors that lead to successful fundraising, a great interest to researchers. This study draws on the social capital theory, human capital theory and level of processing (LOP) theory to predict the success of crowdfunding projects.

A feature set is extracted and correlations between project success and features are utilized to order the features. The artificial neural network (ANN) is popularly applied to analyze the dependencies of the input variables to improve the accuracy of prediction. However, the problem of overfitting may exist in such neural networks. This study proposes a neural network method based on ensemble machine learning and dropout methods to generate several neural networks for preventing the problem of overfitting. Four machine learning techniques are applied and compared for prediction performance.

This study shows that the success of crowdfunding projects can be predicted by measuring and analyzing big data of social media activity, human capital of funders and online project presentation. The ensemble neural network method achieves highest accuracy. The investments rose from early projects and another platform by the funder serve as credible indicators for later investors.

The managerial implication of this study is that the project founders and investors can apply the proposed model to predict the success of crowdfunding projects. This study also identifies the most influential features that affect fundraising outcomes. The project funders can use these features to increase the successful opportunities of crowdfunding project.

This study contributes to apply a new machine learning modeling method to extract features from activity data of crowdfunding platforms and predict crowdfunding project success. In addition, it contributes to the research on the deployment of social capital, human capital and online presentation strategies in a crowdfunding context as well as offers practical implications for project funders and investors.

The aim of this research study is to develop a risk-based framework that can quantify maintainability to forecast future maintainability of a building at early stages as a decision tool to minimize increase of maintenance cost.

A survey-based approach was used to explore the risk factors in the domain of maintainability risks under tropical environmental conditions. The research derived ten risk factors based on 58 identified causes related to maintainability issues as common to high-rise buildings in tropical conditions. Impact of these risk factors was evaluated using an indicator referred to as the “maintenance score (MS)” which was derived from the “whole-life maintenance cost” involved in maintaining the expected “performance” level of the building. Further, an ensemble neural network (ENN) model was developed to model the MS for evaluating maintainability risks in high-rise buildings.

Results showed that predictions from the model were highly compatible and in the same order when compared with calculations based on actual past data. It further showed that, maintainability of buildings could be improved if the building was designed, constructed and managed properly by controlling their maintainability risks.

The ENN model was used to analyze maintainability of a high-rise building. Thus, it provides a useful tool for designers, clients, facilities managers/maintenance managers and users to analyze maintainability risks of buildings at early stages.

Customer churn prediction is one of the most well-known approaches to manage and improve customer retention. Machine learning techniques, especially classification algorithms, are very popular tools to predict the churners. In this paper, three ensemble classifiers are proposed based on bagging and boosting for customer churn prediction.

In this paper, three ensemble classifiers are proposed based on bagging and boosting for customer churn prediction. The first classifier, which is called boosted bagging, uses boosting for each bagging sample. In this approach, before concluding the final results in a bagging algorithm, the authors try to improve the prediction by applying a boosting algorithm for each bootstrap sample. The second proposed ensemble classifier, which is called bagged bagging, combines bagging with itself. In the other words, the authors apply bagging for each sample of bagging algorithm. Finally, the third approach uses bagging of neural network with learning based on a genetic algorithm.

To examine the performance of all proposed ensemble classifiers, they are applied to two datasets. Numerical simulations illustrate that the proposed hybrid approaches outperform the simple bagging and boosting algorithms as well as base classifiers. Especially, bagged bagging provides high accuracy and precision results.

In this paper, three novel ensemble classifiers are proposed based on bagging and boosting for customer churn prediction. Not only the proposed approaches can be applied for customer churn prediction but also can be used for any other binary classification algorithms.

Internet of Things (IoT) is a network, which provides the connection with various physical objects such as smart machines, smart home appliance and so on. The physical objects are allocated with a unique internet address, namely, Internet Protocol, which is used to perform the data broadcasting with the external objects using the internet. The sudden increment in the number of attacks generated by intruders, causes security-related problems in IoT devices while performing the communication. The main purpose of this paper is to develop an effective attack detection to enhance the robustness against the attackers in IoT.

In this research, the lasso regression algorithm is proposed along with ensemble classifier for identifying the IoT attacks. The lasso algorithm is used for the process of feature selection that modeled fewer parameters for the sparse models. The type of regression is analyzed for showing higher levels when certain parts of model selection is needed for parameter elimination. The lasso regression obtains the subset for predictors to lower the prediction error with respect to the quantitative response variable. The lasso does not impose a constraint for modeling the parameters caused the coefficients with some variables shrink as zero. The selected features are classified by using an ensemble classifier, that is important for linear and nonlinear types of data in the dataset, and the models are combined for handling these data types.

The lasso regression with ensemble classifier–based attack classification comprises distributed denial-of-service and Mirai botnet attacks which achieved an improved accuracy of 99.981% than the conventional deep neural network (DNN) methods.

Here, an efficient lasso regression algorithm is developed for extracting the features to perform the network anomaly detection using ensemble classifier.