Entering the Black Box of Neural Networks

2003 ◽  
Vol 42 (03) ◽  
pp. 287-296 ◽  
Author(s):  
B. S. Gerber ◽  
T. G. Tape ◽  
R. S. Wigton ◽  
P. S. Heckerling
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Sensitivity Analysis ◽  
Logistic Regression ◽  
Artificial Neural Networks ◽  
Important Predictor ◽  
Predictor Variables ◽  
Linear Substitution ◽  
Relevance Analysis ◽  
Artificial Neural

Summary Objectives: Artificial neural networks have proved to be accurate predictive instruments in several medical domains, but have been criticized for failing to specify the information upon which their predictions are based. We used methods of relevance analysis and sensitivity analysis to determine the most important predictor variables for a validated neural network for community-acquired pneumonia. Methods: We studied a feed-forward, back-propagation neural network trained to predict pneumonia among patients presenting to an emergency department with fever or respiratory complaints. We used the methods of full retraining, weight elimination, constant substitution, linear substitution, and data permutation to identify a consensus set of important demographic, symptom, sign, and comorbidity predictors that influenced network output for pneumonia. We compared predictors identified by these methods to those identified by a weight propagation analysis based on the matrices of the network, and by logistic regression. Results: Predictors identified by these methods were clinically plausible, and were concordant with those identified by weight analysis, and by logistic regression using the same data. The methods were highly correlated in network error, and led to variable sets with errors below bootstrap 95% confidence intervals for networks with similar numbers of inputs. Scores for variable relevance tended to be higher with methods that precluded network retraining (weight elimination) or that permuted variable values (data permutation), compared with methods that permitted retraining (full retraining) or that approximated its effects (constant and linear substitution). Conclusion: Methods of relevance analysis and sensitivity analysis are useful for identifying important predictor variables used by artificial neural networks.

scholarly journals A State-Based Sensitivity Analysis for Distinguishing the Global Importance of Predictor Variables in Artificial Neural Networks

2016 ◽  
Vol 2016 ◽  
pp. 1-11 ◽  
Author(s):  
Ehsan Ardjmand ◽  
David F. Millie ◽  
Iman Ghalehkhondabi ◽  
William A. Young II ◽  
Gary R. Weckman
Keyword(s):  
Neural Networks ◽  
Sensitivity Analysis ◽  
Artificial Neural Networks ◽  
Network Models ◽  
Predictor Variables ◽  
Apparent Lack ◽  
Artificial Neural ◽  
High Degree ◽  
Universal Approximators ◽  
Empirical Approaches

Artificial neural networks (ANNs) are powerful empirical approaches used to model databases with a high degree of accuracy. Despite their recognition as universal approximators, many practitioners are skeptical about adopting their routine usage due to lack of model transparency. To improve the clarity of model prediction and correct the apparent lack of comprehension, researchers have utilized a variety of methodologies to extract the underlying variable relationships within ANNs, such as sensitivity analysis (SA). The theoretical basis of local SA (that predictors are independent and inputs other than variable of interest remain “fixed” at predefined values) is challenged in global SA, where, in addition to altering the attribute of interest, the remaining predictors are varied concurrently across their respective ranges. Here, a regression-based global methodology, state-based sensitivity analysis (SBSA), is proposed for measuring the importance of predictor variables upon a modeled response within ANNs. SBSA was applied to network models of a synthetic database having a defined structure and exhibiting multicollinearity. SBSA achieved the most accurate portrayal of predictor-response relationships (compared to local SA and Connected Weights Analysis), closely approximating the actual variability of the modeled system. From this, it is anticipated that skepticisms concerning the delineation of predictor influences and their uncertainty domains upon a modeled output within ANNs will be curtailed.

Comparison of Logistic Regression and Artificial Neural Network based Bankruptcy Prediction Models

2015 ◽  
Vol 3 (1) ◽  
Author(s):  
Easwaran Iyer ◽  
Vinod Kumar Murti
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Artificial Neural Network ◽  
Logistic Regression ◽  
Artificial Neural Networks ◽  
Prediction Models ◽  
Bankruptcy Prediction ◽  
Manufacturing Companies ◽  
Recent Developments ◽  
Artificial Neural

Logistic Regression is one of the popular techniques used for bankruptcy prediction and its popularity is attributed due to its robust nature in terms of data characteristics. Recent developments have explored Artificial Neural Networks for bankruptcy prediction. In this study, a paired sample of 174 cases of Indian listed manufacturing companies have been used for building bankruptcy prediction models based on Logistic Regression and Artificial Neural Networks. The time period of study was year 2000 through year 2009. The classification accuracies have been compared for built models and for hold-out sample of 44 paired cases. In analysis and hold-out samples, both the models have shown appreciable classification results, three years prior to bankruptcy. Thus, both the models can be used (by banks, SEBI etc.) for bankruptcy prediction in Indian Context, however, Artificial Neural Network has shown marginal supremacy over Logistic Regression.

scholarly journals A New Hybrid Method Logistic Regression and Feedforward Neural Network for Lung Cancer Data

2012 ◽  
Vol 2012 ◽  
pp. 1-10 ◽  
Author(s):  
Taner Tunç
Keyword(s):  
Neural Network ◽  
Lung Cancer ◽  
Neural Networks ◽  
Artificial Neural Network ◽  
Logistic Regression ◽  
Artificial Neural Networks ◽  
Hybrid Approach ◽  
Cancer Data ◽  
Lung Cancer Data ◽  
Artificial Neural

Logistic regression (LR) is a conventional statistical technique used for data classification problem. Logistic regression is a model-based method, and it uses nonlinear model structure. Another technique used for classification is feedforward artificial neural networks. Feedforward artificial neural network is a data-based method which can model nonlinear models through its activation function. In this study, a hybrid approach of model-based logistic regression technique and data-based artificial neural network was proposed for classification purposes. The proposed approach was applied to lung cancer data, and obtained results were compared. It was seen that the proposed hybrid approach was superior to logistic regression and feedforward artificial neural networks with respect to many criteria.

Predicting Survival in Patients With Brain Metastases Treated With Radiosurgery Using Artificial Neural Networks

Neurosurgery ◽  
2013 ◽  
Vol 72 (6) ◽  
pp. 944-952 ◽  
Author(s):  
Eric K. Oermann ◽  
Marie-Adele S. Kress ◽  
Brian T. Collins ◽  
Sean P. Collins ◽  
David Morris ◽  
...  
Keyword(s):  
Neural Networks ◽  
Sensitivity Analysis ◽  
Logistic Regression ◽  
Artificial Neural Networks ◽  
Primary Tumor ◽  
Performance Status ◽  
Data Sets ◽  
Validation Data ◽  
Clinical Environments ◽  
Artificial Neural

Abstract BACKGROUND: Artificial neural networks (ANNs) excel at analyzing challenging data sets and can be exceptional tools for decision support in clinical environments. The present study pilots the use of ANNs for determining prognosis in neuro-oncology patients. OBJECTIVE: To determine whether ANNs perform better at predicting 1-year survival in a group of patients with brain metastasis compared with traditional predictive tools. METHODS: ANNs were trained on a multi-institutional data set of radiosurgery patients to predict 1-year survival on the basis of several input factors. A single ANN, an ensemble of 5 ANNs, and logistic regression analyses were compared for efficacy. Sensitivity analysis was used to identify important variables in the ANN model. RESULTS: A total of 196 patients were divided up into training, testing, and validation data sets consisting of 98, 49, and 49 patients, respectively. Patients surviving at 1 year tended to be female (P = .001) and of good performance status (P = .01) and to have favorable primary tumor histology (P = .001). The pooled voting of 5 ANNs performed significantly better than the multivariate logistic regression model (P = .02), with areas under the curve of 84% and 75%, respectively. The ensemble also significantly outperformed 2 commonly used prognostic indexes. Primary tumor subtype and performance status were identified on sensitivity analysis to be the most important variables for the ANN. CONCLUSION: ANNs outperform traditional statistical tools and scoring indexes for predicting individual patient prognosis. Their facile implementation, robustness in the presence of missing data, and ability to continuously learn make them excellent choices for use in complicated clinical environments.

scholarly journals Uso de redes neurais artificiais na análise de variáveis ambientais associadas à deposição de serapilheira

2016 ◽  
Vol 7 (3) ◽  
pp. 394
Author(s):  
Priscila Sales Rodrigues Aquino ◽  
Marcos Sales Rodrigues ◽  
Renato Vinícius Oliveira Castro ◽  
Mauro Eloi Nappo
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Sensitivity Analysis ◽  
Artificial Neural Network ◽  
Artificial Neural Networks ◽  
Environmental Variables ◽  
Gallery Forest ◽  
Experimental Site ◽  
Artificial Neural ◽  
X 32

The objective of this study was to evaluate, with the use of artificial neural networks, the influence of some environmental variables in litterfall. The study was conducted on the gallery forest along ‘Lava-pés’ stream in Goiás State, Brazil where the experimental site (3 ha) was structured in a grid of 60 litterfall traps, with 0.33 m2 each and held 0.65 m above the soil, georeferenced and spaced at intervals of 32 x 32 m. Litterfall was monthly collected from December 2011 to November 2012. All litterfall samples were manually separated into three fractions: leaves (LE), branch bark (BB), and reproductive parts (RP). Relevance of climate , temporal, spatial and phytosociological variables in litterfall deposition were evaluated, through sensitivity analysis provided by the artificial neural network with the best performance. According to the statistical analysis, all variables were significant in the phenomenon, while the variable time (months of the year) was the most important for litterfall in the evaluated area. Artificial neural networks are shown as a powerful tool for litterfall analysis.

scholarly journals The Budapest Amyloid Predictor and Its Applications

Biomolecules ◽  
2021 ◽  
Vol 11 (4) ◽  
pp. 500
Author(s):  
László Keresztes ◽  
Evelin Szögi ◽  
Bálint Varga ◽  
Viktor Farkas ◽  
András Perczel ◽  
...  
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Artificial Neural Networks ◽  
Support Vector ◽  
User Needs ◽  
Computational Technique ◽  
Linear Support Vector Machine ◽  
Β Sheets ◽  
Artificial Neural ◽  
Amorphous Aggregation

The amyloid state of proteins is widely studied with relevance to neurology, biochemistry, and biotechnology. In contrast with nearly amorphous aggregation, the amyloid state has a well-defined structure, consisting of parallel and antiparallel β-sheets in a periodically repeated formation. The understanding of the amyloid state is growing with the development of novel molecular imaging tools, like cryogenic electron microscopy. Sequence-based amyloid predictors were developed, mainly using artificial neural networks (ANNs) as the underlying computational technique. From a good neural-network-based predictor, it is a very difficult task to identify the attributes of the input amino acid sequence, which imply the decision of the network. Here, we present a linear Support Vector Machine (SVM)-based predictor for hexapeptides with correctness higher than 84%, i.e., it is at least as good as the best published ANN-based tools. Unlike artificial neural networks, the decisions of the linear SVMs are much easier to analyze and, from a good predictor, we can infer rich biochemical knowledge. In the Budapest Amyloid Predictor webserver the user needs to input a hexapeptide, and the server outputs a prediction for the input plus the 6 × 19 = 114 distance-1 neighbors of the input hexapeptide.

scholarly journals Non-Destructive Micromagnetic Determination of Hardness and Case Hardening Depth Using Linear Regression Analysis and Artificial Neural Networks

Metals ◽  
2020 ◽  
Vol 11 (1) ◽  
pp. 18
Author(s):  
Rahel Jedamski ◽  
Jérémy Epp
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Regression Analysis ◽  
Artificial Neural Networks ◽  
Linear Regression ◽  
Linear Regression Analysis ◽  
Case Hardening ◽  
Artificial Neural ◽  
Non Destructive

Non-destructive determination of workpiece properties after heat treatment is of great interest in the context of quality control in production but also for prevention of damage in subsequent grinding process. Micromagnetic methods offer good possibilities, but must first be calibrated with reference analyses on known states. This work compares the accuracy and reliability of different calibration methods for non-destructive evaluation of carburizing depth and surface hardness of carburized steel. Linear regression analysis is used in comparison with new methods based on artificial neural networks. The comparison shows a slight advantage of neural network method and potential for further optimization of both approaches. The quality of the results can be influenced, among others, by the number of teaching steps for the neural network, whereas more teaching steps does not always lead to an improvement of accuracy for conditions not included in the initial calibration.

scholarly journals Optimal Artificial Neural Network Type Selection Method for Usage in Smart House Systems

Sensors ◽  
2020 ◽  
Vol 21 (1) ◽  
pp. 47
Author(s):  
Vasyl Teslyuk ◽  
Artem Kazarian ◽  
Natalia Kryvinska ◽  
Ivan Tsmots
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Artificial Neural Network ◽  
Artificial Neural Networks ◽  
Input Data ◽  
Optimization Criterion ◽  
Fuzzy Input ◽  
Different Types ◽  
Smart House ◽  
Artificial Neural

In the process of the “smart” house systems work, there is a need to process fuzzy input data. The models based on the artificial neural networks are used to process fuzzy input data from the sensors. However, each artificial neural network has a certain advantage and, with a different accuracy, allows one to process different types of data and generate control signals. To solve this problem, a method of choosing the optimal type of artificial neural network has been proposed. It is based on solving an optimization problem, where the optimization criterion is an error of a certain type of artificial neural network determined to control the corresponding subsystem of a “smart” house. In the process of learning different types of artificial neural networks, the same historical input data are used. The research presents the dependencies between the types of neural networks, the number of inner layers of the artificial neural network, the number of neurons on each inner layer, the error of the settings parameters calculation of the relative expected results.

Performance Estimation of Direct Methanol Fuel Cell Using Artificial Neural Network

2015 ◽  
Author(s):  
M. A. Rafe Biswas ◽  
Melvin D. Robinson
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Artificial Neural Networks ◽  
Fuel Cell ◽  
Direct Methanol Fuel Cell ◽  
Cell Voltage ◽  
Operating Conditions ◽  
Direct Methanol ◽  
Artificial Neural ◽  
Methanol Fuel Cell

A direct methanol fuel cell can convert chemical energy in the form of a liquid fuel into electrical energy to power devices, while simultaneously operating at low temperatures and producing virtually no greenhouse gases. Since the direct methanol fuel cell performance characteristics are inherently nonlinear and complex, it can be postulated that artificial neural networks represent a marked improvement in performance prediction capabilities. Artificial neural networks have long been used as a tool in predictive modeling. In this work, an artificial neural network is employed to predict the performance of a direct methanol fuel cell under various operating conditions. This work on the experimental analysis of a uniquely designed fuel cell and the computational modeling of a unique algorithm has not been found in prior literature outside of the authors and their affiliations. The fuel cell input variables for the performance analysis consist not only of the methanol concentration, fuel cell temperature, and current density, but also the number of cells and anode flow rate. The addition of the two typically unconventional variables allows for a more distinctive model when compared to prior neural network models. The key performance indicator of our neural network model is the cell voltage, which is an average voltage across the stack and ranges from 0 to 0:8V. Experimental studies were carried out using DMFC stacks custom-fabricated, with a membrane electrode assembly consisting of an additional unique liquid barrier layer to minimize water loss through the cathode side to the atmosphere. To determine the best fit of the model to the experimental cell voltage data, the model is trained using two different second order training algorithms: OWO-Newton and Levenberg-Marquardt (LM). The OWO-Newton algorithm has a topology that is slightly different from the topology of the LM algorithm by the employment of bypass weights. It can be concluded that the application of artificial neural networks can rapidly construct a predictive model of the cell voltage for a wide range of operating conditions with an accuracy of 10−3 to 10−4. The results were comparable with existing literature. The added dimensionality of the number of cells provided insight into scalability where the coefficient of the determination of the results for the two multi-cell stacks using LM algorithm were up to 0:9998. The model was also evaluated with empirical data of a single-cell stack.

scholarly journals Artificial neural networks applied for soil class prediction in mountainous landscape of the Serra do Mar¹

2014 ◽  
Vol 38 (6) ◽  
pp. 1681-1693 ◽  
Author(s):  
Braz Calderano Filho ◽  
Helena Polivanov ◽  
César da Silva Chagas ◽  
Waldir de Carvalho Júnior ◽  
Emílio Velloso Barroso ◽  
...  
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Artificial Neural Networks ◽  
Kappa Index ◽  
Class Prediction ◽  
Soil Class ◽  
Serra Do Mar ◽  
The Neural Network ◽  
Artificial Neural ◽  
Local Geology

Soil information is needed for managing the agricultural environment. The aim of this study was to apply artificial neural networks (ANNs) for the prediction of soil classes using orbital remote sensing products, terrain attributes derived from a digital elevation model and local geology information as data sources. This approach to digital soil mapping was evaluated in an area with a high degree of lithologic diversity in the Serra do Mar. The neural network simulator used in this study was JavaNNS and the backpropagation learning algorithm. For soil class prediction, different combinations of the selected discriminant variables were tested: elevation, declivity, aspect, curvature, curvature plan, curvature profile, topographic index, solar radiation, LS topographic factor, local geology information, and clay mineral indices, iron oxides and the normalized difference vegetation index (NDVI) derived from an image of a Landsat-7 Enhanced Thematic Mapper Plus (ETM+) sensor. With the tested sets, best results were obtained when all discriminant variables were associated with geological information (overall accuracy 93.2 - 95.6 %, Kappa index 0.924 - 0.951, for set 13). Excluding the variable profile curvature (set 12), overall accuracy ranged from 93.9 to 95.4 % and the Kappa index from 0.932 to 0.948. The maps based on the neural network classifier were consistent and similar to conventional soil maps drawn for the study area, although with more spatial details. The results show the potential of ANNs for soil class prediction in mountainous areas with lithological diversity.

Sign in / Sign up

Export Citation Format

Share Document