Neural Network-based Decision Research Articles

A neural network-based predictive decision model for customer retention in the telecommunication sector

Acquiring a new customer is far more expensive than retaining a customer. Hence, customer retention is a key aspect of business for a firm to maintain and improve on its market share and profit. The paper analyses customer retention strategies by employing an artificial neural network-based decision model to a real-life dataset collected from 311 mobile service users in India. Seven linear and non-linear adaptive models are developed using features related to customer dissatisfaction (DSF), customer disloyalty (DLF) and customer churn (CF). Findings of this study suggest that non-linear models are most efficient in predicting customer churn, and both DSF and DLF variables significantly affect the retention strategy. Three groups of customers are discussed in this study in the order of least likelihood of churning to most likelihood. Finally, a priority matrix based on key performance indicators is proposed to help service providers target potential customers to retain.

A convolutional neural network-based decision support system for neonatal quiet sleep detection.

Sleep plays an important role in neonatal brain and physical development, making its detection and characterization important for assessing early-stage development. In this study, we propose an automatic and computationally efficient algorithm to detect neonatal quiet sleep (QS) using a convolutional neural network (CNN). Our study used 38-hours of electroencephalography (EEG) recordings, collected from 19 neonates at Fudan Children's Hospital in Shanghai, China (Approval No. (2020) 22). To train and test the CNN, we extracted 12 prominent time and frequency domain features from 9 bipolar EEG channels. The CNN architecture comprised two convolutional layers with pooling and rectified linear unit (ReLU) activation. Additionally, a smoothing filter was applied to hold the sleep stage for 3 minutes. Through performance testing, our proposed method achieved impressive results, with 94.07% accuracy, 89.70% sensitivity, 94.40% specificity, 79.82% F1-score and a 0.74 kappa coefficient when compared to human expert annotations. A notable advantage of our approach is its computational efficiency, with the entire training and testing process requiring only 7.97 seconds. The proposed algorithm has been validated using leave one subject out (LOSO) validation, which demonstrates its consistent performance across a diverse range of neonates. Our findings highlight the potential of our algorithm for real-time neonatal sleep stage classification, offering a fast and cost-effective solution. This research opens avenues for further investigations in early-stage development monitoring and the assessment of neonatal health.

A deep neural network-based decision support system for intelligent geospatial data analysis in intelligent agriculture system

A deep neural network-based decision support system for intelligent geospatial data analysis in intelligent agriculture system

Artificial neural network-based decision support systems in manufacturing processes: A systematic literature review

The use of artificial neural network models to enrich the analytical and predictive capabilities of decision support systems in manufacturing has increased. The growing complexity and uncertainty in the manufacturing sector demand improved decision-making to ensure low operations costs, high productivity, and sustainable use of resources. Artificial neural networks have the inherent capacity to analyze the most uncertain and complex patterns in unstructured decision problems. This review aims to synthesize and provide a comprehensive summary of recent studies on artificial neural network-based decision support systems as applied in manufacturing processes. First, the specific processes in manufacturing where artificial neural network-based decision support systems are used are analyzed. A total of 99 multi-disciplinary publications on artificial neural network-based decision support systems published between 2011 and 2021 are retrieved and processed following a rigorous execution of the designated acceptance criteria and quality assessment. A review of the selected studies indicates a growing interest in applying artificial neural networks in decision support systems. Product and process design, performance evaluation, and predictive maintenance are the main application areas identified. A growing tendency to combine artificial neural network models with other intelligent tools, notably fuzzy logic, and genetic algorithm, is noted to overcome drawbacks such as slow convergence when training the algorithms. Further research should extend to other tools for enriching the performance of artificial neural networks in manufacturing processes.

Neural Network- and Fuzzy Control-Based Energy Optimization for the Switching in Parallel Hybrid Two-Wheeler

Optimization of a two-wheeler hybrid electric vehicle (HEV) is a typical challenge compared to that for four-wheeler HEVs. Some of the challenges which are particular to two-wheeler HEVs are throttle integration, smooth switching between power sources, add-on weight compensation, efficiency improvisation in traffic, and energy optimization. Two power sources need to be synchronized skillfully for optimum energy utilization. A prominent variant of HEV is that it easily converts conventional scooters into parallel hybrids by “Through-the-Road (TTR)” architecture. This paper focuses on three switching control strategies of HEVs based on the state of charge, fuzzy logic, and neural network. Further, to optimize energy usage, all these control strategies are compared. Energy management control for the TTR model is developed with vehicle parameters in the Simulink environment and simulated using the “World Harmonized Motorcycle Test Cycle” (WMTC) drive cycle. The multivariable input model is presented with a fuzzy rule-based hybrid switching control. A similar system is also modeled with a neural network-based decision control and the observations are tabulated for the fuel economy and energy management. Simulation results show that the neural network-based optimization results in minimal energy consumption among all three hybrid operations.

Optimization of a Depiction Procedure for an Artificial Intelligence-Based Network Protection System Using a Genetic Algorithm

The current demand for remote work, remote teaching and video conferencing has brought a surge not only in network traffic, but unfortunately, in the number of attacks as well. Having reliable, safe and secure functionality of various network services has never been more important. Another serious phenomenon that is apparent these days and that must not be discounted is the growing use of artificial intelligence techniques for carrying out network attacks. To combat these attacks, effective protection methods must also utilize artificial intelligence. Hence, we are introducing a specific neural network-based decision procedure that can be considered for application in any flow characteristic-based network-traffic-handling controller. This decision procedure is based on a convolutional neural network that processes the incoming flow characteristics and provides a decision; the procedure can be understood as a firewall rule. The main advantage of this decision procedure is its depiction process, which has the ability to transform the incoming flow characteristics into a graphical structure. Graphical structures are regarded as very efficient data structures for processing by convolutional neural networks. This article’s main contribution consists of the development and improvement of the depiction process using a genetic algorithm. The results presented at the end of the article show that the decision procedure using an optimized depiction process brings significant improvements in comparison to previous experiments.

Neural network-based blockchain decision scheme

ABSTRACT Blockchain is a game-changer for many applications such as storage, consensus, authentication, and many others. Although blockchain’s contribution is undeniable, but it is not always the best solution. Professionals struggle to decide whether to adopt blockchain and when they do, they are faced with another challenge to choose the suitable type. Decision schemes are developed to help professionals select the best technology. Existing decision schemes are of two types questionnaire forms and flow diagrams. All existing schemes share a main drawback that the user is limited to a predefined set of answers without having the capacity to fine-grain his preferences in addition to per scheme-specific drawbacks. In this paper, we will propose a neural network-based decision scheme that gives the user an advanced decision support feature of specifying proportional weights between characteristics that does not exist in any other blockchain type decision scheme and solve the drawbacks available in existing schemes.

Machine learning and decision support system on credit scoring

Among the numerous alternatives used in the world of risk balance, it highlights the provision of guarantees in the formalization of credit agreements. The objective of this paper is to compare the achievement of fuzzy sets with that of artificial neural network-based decision trees on credit scoring to predict the recovered value using a sample of 1890 borrowers. Comparing with fuzzy logic, the decision analytic approach can more easily present the outcomes of the analysis. On the other hand, fuzzy logic makes some implicit assumptions that may make it even harder for credit-grantors to follow the logical decision-making process. This paper leads an initial study of collateral as a variable in the calculation of the credit scoring. The study concludes that the two models make modelling of uncertainty in the credit scoring process possible. Although more difficult to implement, fuzzy logic is more accurate for modelling the uncertainty. However, the decision tree model is more favourable to the presentation of the problem.

Bimodal classification of English allophones employing acoustic speech signal and facial motion capture

A method for automatic transcription of English speech into International Phonetic Alphabet (IPA) system is developed and studied. The principal objective of the study is to evaluate to what extent the visual data related to lip reading can enhance recognition accuracy of the transcription of English consonantal and vocalic allophones. To this end, motion capture markers were placed on the faces of seven speakers to obtain lip tracking data synchronized with the audio signal. 32 markers were used, 20 of which were placed on the speaker's inner lips and 4 on a special cap, which served as the point of reference and stabilized the FMC image while post-processing. Speech samples were simultaneously recorded as a list of approximately 300 words in which all English consonantal and vocalic allophones were represented. Different parameterization strategies were tested and the accuracy of vocalic segments recognition in different experiments was analyzed. The process of multimodal feature extraction is explained, the applied feature selection methods are presented and the obtained results are discussed. Further challenges related to the bi-modal feature extraction process and neural network-based decision systems employment are discussed. [Research sponsored by the Polish National Science Centre, Dec. No. 2015/17/B/ST6/01874.]

Artificial Neural Network-Based Decision Support System for Development of an Energy-Efficient Built Environment

Implementing energy-efficient solutions in a built environment is important for reaching international energy reduction targets. For advanced energy efficiency-related solutions, computer-based decision support systems are proposed and rapidly used in a variety of spheres relevant to a built environment. Present research proposes a novel artificial neural network-based decision support system for development of an energy-efficient built environment. The system was developed by integrating methods of the multiple criteria evaluation and multivariant design, determination of project utility and market value, and visual data mining by artificial neural networks. It enables a user to compose up to 100,000,000 combinations of the energy-efficient solutions, analyze strengths and weaknesses of a built environment projects, provide advice for stakeholders, and calculate market value and utility degree of the projects. For visual data mining, self-organizing maps (type neural networks) are used, which may influence the choosing of the final set of alternatives and criteria in the decision-making problem, taking into account the discovered similarities of alternatives or criteria. A system was validated by the real case study on the design of an energy-efficient individual house.

Development of a decision support system to predict physicians’ rehabilitation protocols for patients with knee osteoarthritis

To design a medical decision support system (MDSS) that would accurately predict the rehabilitation protocols prescribed by the physicians for patients with knee osteoarthritis (OA) using only their demographic and clinical characteristics. The demographic and clinical variables for 170 patients receiving one of three treatment protocols for knee OA were entered into the MDSS. Demographic variables in the model were age and sex. Clinical variables entered into the model were height, weight, BMI, affected side, severity of knee OA, and severity of pain. All patients in the study received one of three treatment protocols for patients with knee OA: (a) hot packs, followed by electrotherapy and exercise, (b) ice packs, followed by ultrasound and exercise and (c) exercise alone. The resilient back propagation artificial neural network algorithm was used, with a ten-fold cross-validation. It was estimated that the MDSS is able to accurately predict the treatment prescribed by the physician for 87% of the patients. We developed an artificial neural network-based decision support system that can viably aid physicians in determining which treatment protocol would best match the anthropometric and clinical characteristics of patients with knee OA.

MLP neural network-based decision for power transformers fault diagnosis using an improved combination of Rogers and Doernenburg ratios DGA

Dissolved gas analysis (DGA) is a widely-used method to detect the power transformer faults, because of its high sensitivity to small amount of electrical faults. The DGA is exploited for fault classification tools implementation using the artificial intelligence techniques. In this study, we use the Rogers ratios, the Doernenburg ratios methods and our proposed combination of Rogers and Doernenburg ratios DGA methods as gas signature. The multi-layer perceptron neural network (MLPNN) is applied for decision making. The paper presents a comparative study on one hand for the choice the most appropriate DGA method and to resolve the problem of conflict between the Rogers and Doernenburg ratios methods. On the other hand, it compares the various MLP architectures by comparing two output data types and three hidden layer types with the aim to establish the most appropriate MLP model. Before testing, the proposed structures are trained and tested by the experimental data from Tunisian Company of Electricity and Gas (STEG). The test results suggest that MLPNN ratios combination can generalize better than other MLPNN models. The approach has the advantages of high accuracy. The other advantage is that the model is practically applicable and may be utilized for an automated power transformer diagnosis. The classification accuracies of the MLPNN classifier are compared with fuzzy logic (FL), radial basis function (RBF), K-nearest neighbor (KNN) and probabilistic neural network (PNN) classifiers. The test results indicate that the developed preprocessing approach can significantly improve the diagnosis accuracies for power transformer fault classification.

A model for bundling mobile value added services using neural networks

Mobile value added services has emerged as one of the significant revenue generators for the mobile telecommunication industry. With the large array of service offerings, telecom providers engage in active advertising of service bundles to attract new as well as retain the existing customers. Focused advertising by personalising the service bundles based on individual customer preferences is a key success factor of such campaigns. This paper proposes the application of a neural network-based decision support model for pattern association using which mixed bundles of such services may be identified, as per customer preferences, so as to induce higher conversion in cross-sales and up-sales. Consumer service preferences can be combined with some standard packages of VAS bundles to personalise the service bundles and these service bundles would be generated dynamically. Identifying and promoting such services identified through pattern association rules will have a direct impact on the overall profitability of the service provider.

SCFNN-Based Decision Feedback Equalization Robust to Frequency Offset and Phase Noise

This paper proposes a self-constructing fuzzy neural network-based decision feedback equalizer (SCFNN DFE). An online learning algorithm containing the structure and parameter learning phases is employed in training the SCFNN DFE. Specifically, the feedforward input vector classification and a gradient-descent method are both used in this online learning algorithm. We show by simulations that the proposed SCFNN DFE offers improvement compared to the traditional DFE methods in the presence of frequency offset and phase noise.

Fast self-constructing fuzzy neural network-based decision feedback equaliser in time-invariant and time-varying channels

A fast self-constructing fuzzy neural network-based decision feedback equaliser (FSCFNN DFE) is proposed. Without estimating the channel, a fast learning algorithm containing the structure and parameter learning phases is employed to the FSCFNN DFE. Both the partition of the feedforward input space and the gradient descent method are used simultaneously with the aid of decision feedback inputs in this fast learning procedure. The performance of FSCFNN DFE is compared with traditional non-linear equalisers in both time-invariant and time-varying channels. The reduced complexity and excellent performance of the FSCFNN DFE make it suitable for severely distorted channel equalisation.

Induction machine fault detection using stray flux EMF measurement and neural network-based decision

The aim of this paper is to present the performances of voltage unbalance and rotor fault detections using an external stray flux sensor in a working three-phase induction machine. The automatic classification and fault severity degree evaluation are realized by using a neural network approach based on a multi-layer perceptron (MLP) structure. In this paper, it is proved that a simple external stray flux sensor is more efficient than the classical stator current sensor to detect rotor broken bar and voltage unbalance, using data processing at low-frequency resolution.

Dynamic decision support graph—Visualization of ANN-generated diagnostic indications of pathological conditions developing over time

A common objection to using artificial neural networks in clinical decision support systems is that the reasoning behind diagnostic indications cannot be sufficiently well explained. This paper presents a method for visualizing diagnostic indications generated from an artificial neural network-based decision support algorithm (ANN-algorithm) in conditions developing over time. The main idea behind the method is first to calculate and graphically present the decision regions corresponding to the diagnostic indications given as output from the ANN-algorithm, in the space of two selected, clinically established 'display variables'. Secondly, the trajectory of time series measurement results of these, often biochemical markers, together with the respective 95% confidence intervals are superimposed on the decision regions. This will permit a nurse or clinician to grasp the diagnostic indication graphically at a glance. The indication is further presented in relation to clinical variables that the clinician is already familiar with, thus providing a sort of explanation. The predictive value of the indication is expressed by the proximity of the measurement result to the decision boundary, separating the decision regions, and by a numerically calculated individualized predictive value. The method is illustrated as applied to a previously published ANN-algorithm for the early ruling-in and ruling-out of acute myocardial infarction, using monitoring of measurement results of myoglobin and troponin-I in plasma. The method is appropriate when there is a limited number of clinically established variables, i.e. variables which the clinician is used to taking into account in clinical reasoning.

A binary neural decision table classifier

In this paper, we introduce a neural network-based decision table algorithm. We focus on the implementation details of the decision table algorithm when it is constructed using the neural network. Decision tables are simple supervised classifiers which, Kohavi demonstrated, can outperform state-of-the-art classifiers such as C4.5. We couple this power with the efficiency and flexibility of a binary associative-memory neural network. Initially, we demonstrate how the binary associative-memory neural network can form the decision table index to map between attribute values and data records and subsequently we show how two attribute selection algorithms can be used to pre-select attributes for this decision table. The attribute selection algorithms are easily implemented within the same binary associative-memory framework producing a tightly coupled, two-tier system allowing attribute selection and decision table indexing. The first attribute selector uses mutual information between attributes and classes to select the attributes that classify best. The second attribute selector uses a probabilistic approach to evaluate randomly selected attribute subsets.

A Web-based Telehealth System for Medical Diagnosis and Information Management

abstractIn this paper, the development of a web-based telehealth system for medical diagnosis and information management is described. Known as “Online eMEDSYS”, the system is designed as a prototype tool for the early diagnosis of Acute Coronary Syndrome (ACS). There are two main modules in the system: Online Medical Database System (OMDS) and Online Decision Support System (ODSS). The OMDS is used for maintaining and administrating medical data and information electronically, while the ODSS is an artificial neural network-based intelligent decision support tool for assisting medical practitioners to make accurate and early diagnosis of ACS. The overall effectiveness of the system is assessed using a real database of suspected ACS patients. Various operating strategies of the system in performing medical diagnosis are demonstrated. The results are analysed, and implications of the system as a useful and usable telehealth tool are discussed.

Transferability of neural network-based decision support algorithms for early assessment of chest-pain patients

The present investigation concerns methodological and epidemiological aspects of the transferability of artificial neural network-based algorithms, as key-components for classification in decision support systems (DSS). The prevalence of pathological conditions to be detected must be known in order to tune an artificial neural networks (ANN)-decision algorithm so that the predictive values of the outcome fulfil medical requirements. Another aspect of transferability, when clinical laboratory results are used, concerns differences in analytical performance of measuring instruments. The relative bias between two instruments is not known exactly, but must be estimated and corrected for. A general method, based on original measured data sets and statistical modeling, was developed for simulating the impact of various correction procedures when using different analytical instruments. The simulation methodology was applied to a real clinical problem of ruling-in/ruling-out of patients with suspected acute myocardial infarction (AMI) by biochemical monitoring. The recommended correction procedure was based on method comparison with use of five duplicate measurements on a common set of patient samples covering the relevant measuring interval. Transferability of laboratory data over time was also studied. The design of quality assurance procedures should be based on analytical quality requirement specifications related to medical needs. Limits of critically sized systematic errors were assessed by calculating the decrease in diagnostic performance of the ANN-algorithm as a result of temporary analytical disturbances. The consequences for the design of QA procedures was illustrated. It is concluded that the actual ANN-decision algorithm for early assessment of chest-pain patients should be possible to transfer to new sites under realistic conditions.