Imad A. Basheer

Computational neural networks for predictive microbiology II. Application to microbial growth

The growth of a specific microorganism on a certain food is influenced by a number of environmental factors such as temperature, pH, and salt concentration. Methods that delineate the history of the growth of microorganisms are always subject to a considerable debate and scrutiny in the field of predictive microbiology. Regardless of its types, a growth model (e.g., modified Gompertz model) contains several parameters that vary depending on the microorganisms/food combination and the associated prevailing environmental conditions. The growth model parameters for a set of operating conditions are commonly determined from expressions developed via multiple linear regressions. In the present study, a substitute for the nonlinear regression-based equations is developed using computational neural networks. Computational neural networks are applied herein on experimental data pertaining to the anaerobic growth of Shigella flexneri. Results have indicated that predictions by neural networks offer better agreement with experimental data as compared to predictions obtained via corresponding regression equations.

On the identification of compaction characteristics by neuronets

Abstract Compaction of soils is aimed at modifying their engineering properties to fulfil the needs of earthwork projects. The two characteristic compaction parameters, namely the optimum moisture content and the maximum dry density, can only be determined experimentally. In this paper, neuronets have been developed and used to determine these two parameters from variables pertinent to geotechnical engineering soil properties and indices. The predictions of the various developed neuronets are compared with corresponding actual values and values obtained via statistical correlation models. Moreover, the advantages of neuronets as prediction techniques over the conventional regression methods are addressed.

Computational neural networks for predictive microbiology: I. methodology

Artificial neural networks are mathematical tools inspired by what is known about the physical structure and mechanism of the biological cognition and learning. Neural networks have attracted considerable attention due to their efficacy to model wide spectrum of challenging problems. In this paper, we present one of the most popular networks, the backpropagation, and discuss its learning algorithm and analyze several issues necessary for designating optimal networks that can generalize after being trained on examples. As an application in the area of predictive microbiology, modeling of microorganism growth by neural networks will be presented in a second paper of this series.

Neural Modeling of Kansas Soil Swelling

Damage due to soil swelling is very noticeable in a wide spectrum of structures such as roads, buildings, canal linings, and landfill liners. To control or overcome any damage, swelling soils are commonly chemically stabilized (e.g., by lime treatment). To evaluate the severity of swelling and design for the best and most economical stabilization strategy, an accurate assessment of the swell potential is acquired. In this study, a huge data base representing 413 soils retrieved from 45 different projects covering 28 counties in Kansas was used to develop prediction models. Neural network-based models and a statistical model were developed. It is shown that neural models provide significant improvements in prediction accuracy over statistical models.

Swelling Potential of Kansas Soils: Modeling and Validation Using Artificial Neural Network Reliability Approach

Determination of the swell potential of a troublesome soil can only be made possible through the use of systematic methods for identifying, testing, and evaluating that potential. For proper evaluation of the severity of swelling for such soils, an accurate soil swell potential assessment method is warranted. To address this key issue, a two-phase research study was performed to develop combined artificial neural network and reliability-based soil swell prediction models. In Phase 1, a reasonable-sized database representing 514 swell soil tests retrieved from over 51 different projects in Kansas was used to develop both neural network–based (NNB) and statistical-based (SB) swell potential prediction models. Direct comparison of results obtained showed that NNB models provide significant improvements in prediction accuracy over their SB counterparts. In the second phase, predictions obtained using the developed NNB models along with the available experimental database were used to produce reliability (probability) factor matrices, which are used to assign a specific confidence level to predictions obtained via NNB models in order to classify the soil under consideration as a swelling or nonswelling type.

Reliability-Based Design of Reinforced Earth Retaining Walls

Reliability of an earth structure can be assessed from the knowledge of the governing probability distribution and its related statistical parameters, namely, the mean and variance. In this study, the mean and variance for the design parameters (width and length of the reinforcing ties) of reinforced earth retaining walls supporting sandy soils are determined using the first-order Taylor series approximation. Design diagrams that enable estimation of both mean and variance also are developed to avoid extensive computations that involve partial differentiation. Errors associated with truncating second-order terms are also evaluated. It is found that for soils with moderately variable physical parameters, the first-order approximation is adequate for estimating both the mean and variance.

Neuronet-based approach to modeling the durability of aggregate in concrete pavement construction

The durability of aggregate used in concrete pavements construction is commonly assessed by subjecting small concrete beams containing the aggregate to cyclic freezing and thawing. The durability of aggregate and concrete specimens is quantified by measuring the durability factor (DF) and percent expansion (EXP). A typical durability test may last 3 to 5 months and involve high costs. It was assumed that the durability of aggregate used as a constituent in concrete elements may be related to some easily measured physical properties of the aggregate. A data base obtained from records of the Kansas Department of Transportation contained a total of 750 durability tests. The observed wide scatter in the experimental data when DF or EXP is related to one physical parameter suggested the use of artificial neural networks to model durability. Neural network models were developed to predict durability of aggregate from five basic physical properties of the aggregate. The models were found to classify the aggregates with regard to their durability with a relatively high accuracy. In addition, the models were used to assess the reliability of prediction. To illustrate the use of the models, numerical examples are presented.

Development of a Universal Cracking Amount Model for Fatigue and Reflective Cracking of Asphalt Pavements

AbstractThis paper reviewed mechanistic-empirical (ME) design and analysis frameworks and identified a weak link: a cracking amount model connecting pavement damage (predicted from mechanistic-empi...