Spectroscopy-Based Partial Prediction of In Vitro Dissolution Profile Using Artificial Neural Networks

This report with recommendations is the result of an expert panel meeting on PAT applications in food industry that was organized by the M3C Section of the European Society of Biochemical Engineering Science (ESBES) at the 10th ESBES Symposium. The aim of the panel was to provide an update on the present status of the subject and to identify critical needs and issues for wider applications of PAT in food industry. A brief description of the current state-of-the-art and industrial uptake of the methodology is provided in this report. It concludes with a number of recommendations to facilitate further developments and a wider application of PAT in food industry. Process Analytical Technologies (PAT) [1] (European Medicines Agency EMA-FDA pilot program for parallel assessment of Quality by Design applications; Document EMA/172347/2011) have been extensively discussed in literature, particularly with respect to (bio)pharmaceutical process modelling, monitoring and control [2]. Table 1 provides an historic overview of the PAT development in the context of food applications. Although successful applications within food industry are increasingly being reported [3, 4], the session on PAT in food industries at the ESBES-IFIBiop 2014 in Lille highlighted significant challenges and opportunities for further development in this area. This position paper briefly reviews the current state-of-the-art, industrial needs and opportunities as well as scientific challenges to be addressed in order to extend the use of PAT methodology in the food industries. Currently, quality and safety control are still based mainly on discontinuous analysis with traditional analytical methods in the lab or, at best, at-line measurements. This is no longer sufficient to fulfill the needs of the food industry. Due to higher safety and quality standards and demands and high throughput of production facilities, the number of samples to be analyzed is increasing. Rapid analysis methods and PAT are required to address these needs along the complete production chain leading to a better understanding and control of raw materials, intermediate products in the production process as well as the final products to be packaged and delivered. The goal is to achieve real time analysis in order: • to avoid usage of any out-of-specification raw material and to detect adulteration, substitution, tampering and counterfeiting leading to non-authentic products; • to be able to intervene and stop/change processes in order to secure the target quality; • to assess the final quality to avoid out-of-specification products being packaged and shipped, thus leading to undesirable customer dissatisfaction and costs associated with the resolution of complaints. Over a hundred years ago (on May 22nd, 1913) the first patent on a PAT device ("Pfeiffenanalysator" for measuring the ratio of H2 and N2 gas for ammonia synthesis) was granted to Paul Gmelin from Badische Anilin- und Soda-Fabrik, BASF (Patentschrift Nr. 281157, Klasse 42/. Gruppe 4). Since then, PAT found broad application in chemical industry, which is dominated by highly automated, continuous processes. Today in chemical processes, such as the synthesis of isocyanates, e.g. hexamethylendiisocyanate (HDI), typically 60-130 PAT measurements are collected. In contrast, in sectors like (bio)pharmaceutical or food industry, the application of PAT is significantly less frequent today, especially, in terms of on-line analytics, where the measurement takes place in or close to the production step (or unit operation). One reason for this is the reduced degree of automation of the processes in these industries, which are dominated by unit operations, single production steps consecutively executed after each other, resulting in production batches. Analytics of a production batch takes place mostly off-line in analytical laboratories during hold-up times of the process intermediate between the different unit operations. Regulatory requirements represent an equally important driver for extensive online analytics. With the 2004 FDA's Quality by Design initiative, regulatory agencies demand systematic risk based process development and understanding of the processing space. This also allows for flexible process adjustments within the explored space for producing the desired product quality. In order to achieve the required understanding, more extensive analytical data is required, ideally obtained in-time on raw materials, process intermediates and ideally also on product quality attributes. In food industry, where typically high product titers are obtained, product analytics is sometimes possible by in-line measurements, such as spectroscopy (section 4). In biopharmaceutical processes, product titers are usually very low, the products are proteins with high molecular weight, and quality attributes are challenging to analyse. In these situations, online sampling with automated sample preparation and analytics may open up a solution [5]. But, identifying and implementing analytical methods, which are fast enough for delivering results on product quality attributes (e.g. protein glycosylation) in time, during the unit operation for allowing process modulation, remains a challenge. The overall typical goals in industry are high safety, high and stable product quality, high yield, low consumption of resources (materials, energy, room, time, and people), reduced influence of variability in raw materials as well as an increased shelf life of products which holds also in food processes. However, compared to the chemical and biopharmaceutical industries, food industry has to deal with certain characteristics, which renders PAT a formidable challenge in this context. The characteristic features in food industry are: • Raw materials are not pure substances: they are complex combinations of pure substances with varying compositions • Raw materials are soft, variable size, fragile and slippery • Physical properties of raw materials depend heavily on temperature, pressure, moisture and harvesting, and storage conditions • Raw materials undergo usually a phase transformation during processing and hence change their physical properties during processing • Micro processes (physical, (bio)-chemical, microbiological) are frequently not known • Highly perishable products. Here the challenge is to produce good quality, which maintains for a long time (shelf life) • High demand for hygiene These characteristics lead to a challenge for the application of sensors in the food industry. For important quality and process variables, such as sensory assessment, micro flora or spoilage, reliable and robust sensors are not yet available. Moreover, the sensors which are available are typically used only in isolated applications and frequently provide insufficient reliability. They are often not integrated in a common data management infrastructure. The materials of construction do not always consider suitability of contact with food and the solutions developed within advanced research projects frequently demand high care and maintenance. One reason why PAT is not as common as in other areas is that in food industry knowledge from various subject areas is necessary (such as physics, chemistry, biology, mathematics, informatics, engineering, nutritional science). For unit operations mathematical models are available in principle, but frequently they are too complicated to adapt. One of the biggest challenges in food industries is the dynamic nature of the processes. Changes in geometry, porosity, microstructure, solubility as well as mass and energy transfer coefficients must be addressed. Mostly gradients of temperature and moisture have to be considered as well as changes in kinetic parameters during process run. End product qualities like color, smell and 'brokens' also strongly influence the process. Therefore, food processes can be considered to be significantly more complex then chemical and bio-pharmaceutical processes. To achieve more stable processes PAT applications are necessary to check the quality of the raw and processed materials and their relationship to each other. If the quality is changing, then control actions resulting in online parameter changes are required in order to maintain constant product quality. Measurement systems are required, which guarantee that the process is in accordance with recipe and formulation. Furthermore, disturbances must be compensated for using control actions to reduce process variability and to conform to food regulations. As Glassey [6] argues, PAT methodology can aid in product design and testing as well as in ensuring full compliance with the HACCP and ISO 22000:2005 requirements during processing. As demonstrated below, PAT methodologies have the potential to aid the identification of critical control points and their critical limits, their effective monitoring and control, but also effective communication with suppliers and customers. Documentation which requires measurements of important variables is especially important so that traceability can be guaranteed. Here PAT has to deliver the corresponding measurement systems. Clearly a high demand for PAT in food industry is evident as is the need for further developments in the science and technology that help address the specific challenges posed by the characteristics of the food industry highlighted above. The complexity of raw materials that are typically soft and easily damageable is arguably one of the major scientific and technological challenges in food industry. During storage and processing the quality of such raw materials can decline due to oxidation processes, pressure and temperature effects. In addition, the visual impression of the final product (i.e. its appearance) is much more important than in other industries due to the fact that this will influence the purchase decision of the consumer. During the whole processing from raw material to the storage of the final product, hygiene is of utmost importance. Therefore non-invasive sensor systems are required in these applications. Measurement systems based on optical principles are particularly suitable from this perspective. Such measurement systems, including near infrared (NIR), Raman and fluorescence spectroscopy as well as computer-assisted image-based systems, potentially have a number of advantages in food process supervision and automation. Spectroscopic methods and imaging devices are well suited for PAT purposes because they are fast, non-destructive, provide multiple chemical information, allowing remote in-process analysis via fiber optics or instruments mounted directly on-line. Fluorescence spectroscopy is the most sensitive spectroscopic technique. Recently, many applications have been developed using fluorescence techniques [3, 7, 8]. Raw materials, the supervision of processing as well as product quality and the contamination of the equipment can be monitored by fluorescence. For example, Everard et al. [3] presented a method for detection of fecal contamination on spinach leaves. They coupled three hyperspectral imaging (HSI) configurations with two multivariate image analysis techniques and compared fluorescence imaging in the visible region with ultra violet and violet excitation sources, and reflectance imaging in the visible to near-infrared regions. They showed, that both fluorescence configurations had 100% detection rates for fecal contamination up to 1:10 dilution level and violet HSI had 99% and 87% detection rates for 1:20 and 1:30 levels, respectively. Everard et al. emphasized that on-line detection of fecal contamination on leaves has the potential to reduce the cases of food borne illnesses and their associated costs. A similar approach is presented by Lee et al. [7] where bovine faeces on Romaine lettuce and baby spinach leaves were investigated. They pointed out that two-band ratios using bands at 665.6 nm and 680.0 nm for lettuce and at 660.8 nm and 680.0 nm for spinach effectively differentiated all contamination spots applied. Grote et al. [8] described a fluorescence measurement technique to monitor a sourdough fermentation process. For the prediction of pH value and acidity during rye sourdough fermentations they applied partial least squares regression and principal component regression models for prediction and compared them with an evaluation where principal component analysis was combined with artificial neural networks. Depending on process operation and evaluation technique the average percentage root mean square errors of prediction for pH values were between 2.5 and 5.1%. For the prediction of the acidity level, the best results were between 6.0 and 8.1%. Liu et al. [9] used the Hoffman reaction to convert acrylamide to a compound which shows strong fluorescence emission at 480 nm. They showed good correlation of acrylamide in the range of 0.015 μg/mL to 20 μg/mL. Using this technique the food security will be increased. A fluorescence imaging device to detect deli residues on deli slicers were used by Beck et al. [10] processing four cheeses and four processed meats. The authors suggested that a fluorescence imaging device can be applied for routine use even in delicatessens. Fig. 1 clearly shows that the application of fluorescence for the monitoring of food processes increased from less than 10 before 2003 to more than 60 a year since 2013. This figure includes all papers from a search containing the key words "Food" and "Fluorescence". Although most of these papers describe laboratory applications rather than industrial PAT applications, they indicate potential future applications in this technology in food industry. Number of published papers obtained from the Scopus database searching for the words "Food" and "Image analysis" or "Raman" or "Near infrared spectroscopy" or "Fluorescence spectroscopy". Raman spectroscopy can only be applied if no fluorescence occurs in the corresponding excitation range. He et al. [11] applied the surface enhanced Raman scattering spectroscopy to detect banned food additives, such as Sudan I dye and Rhodamine B in food, Malachite green residues in aquaculture fish. They concluded, that Raman spectroscopy and chemometric evaluation techniques can be used to identify banned food additives to ensure food safety. Ilaslan et al. [12] presented a method based on Raman spectroscopy to provide a rapid method for evaluating the quantitative analysis of glucose, fructose, and sucrose in soft drinks. Wang et al. [13] applied a Raman spectrometer as a process analyzer to monitor the wine fermentation. They demonstrated that sugar, ethanol and glycerol can be measured on-line with high correlation (higher than 0.98) to the HPLC reference measurements. Nache et al. [14] investigated Raman spectra from pork meat to monitor the early postmortem lactate accumulation and pH decline. They suggested that the locally weighted regression applied to the standard normal variate (SNV) normalized Raman spectra provide one of the most accurate and robust models with a cross-validated coefficient of determination (r2cv) of 0.97 for pH and lactate, a cross-validated root mean square error (RMSECV) of 4.5 mmol/kg for the lactate prediction and 0.06 pH-units for the pH prediction. These results demonstrate significant potential of combining chemometrics and Raman spectroscopy for on-line meat quality control applications. Fig. 1 shows the number of papers recorded in Scopus (search words "Food" and "Raman"). Compared to fluorescence, the number of papers describing Raman spectroscopy is twice as high in recent years with a significant rate of increase, demonstrating the increasing interest in this method for food process monitoring. Computer vision systems enable one of the main aspects of consumer preference – the appearance of a product – to be inherently considered. Therefore, computer vision systems for the supervision of food processing also gained importance over the years. An overview of several examples is given by Sun [15]. Especially the supervision of food drying processes is discussed by Aghbashlo at al. [16]. They pointed out that there is a large unexploited potential in the image data captured during various food processing operations. More informative feature extraction algorithms or novel pattern recognition procedures have to be developed. Paquet-Durand et al. [4] described a system for the supervision of the baking process. They demonstrated, by using the Viola–Jones-algorithm as well as neural networks, that the pastry can be identified and the volume increase as well as color development can be monitored. To detect defective apples Zhang et al. [17] used a computer vision system, which was combined with an automatic lightness correction system. For 160 samples they showed a 95% overall detection accuracy. For the evaluation they used a weighted relevance vector machine classifier. Fig. 1 shows the number of papers over time from Scopus searching for "Food" and "Image analysis". Here the applications start earlier compared to the spectroscopic methods, the number of applications per year is higher than in the case of spectroscopic applications; except in the year 2014, where more reports of Raman applications were published. Moreover imaging systems using visible and NIR region are available to perform quality checks based on the spectroscopic information derived from each point in the image. One example is the on-line analysis of the widely varying fat distribution in salmon for sorting purposes [18]. Near infrared (NIR) is a well-established method for rapid analysis of food raw materials and products [18], either on-/in-line, at-line or off-line. In the lab or at-line multiple components such as fat, protein, moisture and many more, can be analyzed without any sample preparation in all kinds of liquid, solid and semi-solid samples. Raw materials can be verified for identity and further characterized regarding composition already in the goods reception. The composition, freshness and adulteration of edible oils [19], e.g. olive oil [20], can be analyzed rapidly before a truck is unloaded. Many other sample types like meat, grains, flour, dairy products [21] and others can be analyzed. Even mixtures like vitamin premixes can be analyzed to make sure that the correct material was delivered and can be used in production. There are several approaches and technologies for PAT available, ranging from simple filter based devices over dispersive diode array spectrometers to Fourier-Transform (FT)-NIR instruments. Samples can be analyzed in-line by fiber optic probes or by contactless systems (Fig. 2). Smaller instruments can be directly attached to pipes, chutes or in other installations. A) NIR reflection probe in a fluid bed dryer. (B) Probe head for contactless measurement of grain and other solids. Applications cover a broad range from a simple monitoring of moisture content of a product on a conveyor belt to more complex control situations. In dairy industry there are demands to control milk powder production by monitoring the feed of the spray tower and the powder after fluid bed drying. Other important production steps to control are the standardization of milk, cream, whey and concentrates [22]. In frying processes with big volumes of oil an in-line monitoring of the oil quality with regards to acid value, anisidine value and the content of polar and polymerized components [23] are of particular interest. Another example of high volume processes where monitoring of protein, moisture and ash is important is the milling of grain to flour and the production of cereals of any kind. Finally there are more and more fermentation processes controlled by in-line NIR to optimize the conditions and follow the feeding and consumption cycles during biomass buildup [24]. In 2013 roughly 120 papers were published, however the recent increase is not as steep as in the case of Raman spectroscopy (Fig. 1). The economic benefits due to PAT can be attributed to higher product quality and yield as well as decreased product variation. This will reduce the overall production costs significantly and increase competitiveness. Furthermore, the knowledge and understanding of the process will increase. However, the level of automation in industrial food processes is significantly lower compared to the chemical and pharmaceutical processes. This is partly due to the complexity of food processes. On the other hand, the lack of reliable sensor systems to determine important process parameters and variables contributed to this in the past. Sensors represent a fundamentalpart of all automation systems, although they only represent a part of the requirements for wider application of PAT in food industry. The areas of further development and recommendations enabling more extensive use of the PAT within this industry also include: • Further industrial case studies and wider dissemination of the positive impact of reliable sensor technology in raw material and product quality as well as process monitoring. • Operational and maintenance requirements of sensor technologies will remain a significant aspect of future sensor technology uptake in this industry. • Data management and analysis software will play a significant role in extending the PAT application within the food industry. Platform solutions, preferably supporting wireless data transfer, at competitive pricing levels are required. • Cost benefit analysis studies are particularly important given the business drivers of this industry. • The impact of PAT is likely to be more pronounced in the new processes, although it is important to continue to encourage the application of PAT approaches within established food processes to enhance the process economics and compliance with the standards ensuring safe food supply chains. • PAT can significantly contribute to the stringent traceability and documentation requirements, although issues of compatibility and data standardization will impact upon data management systems employed by the companies within the supply chain. The authors declare no financial or commercial conflict of interest.

The generalization ability of artificial neural networks (ANNs) and M5 model tree (M5Tree) in modeling reference evapotranspiration (ET 0 ) is investigated in this study. Daily climatic data, average temperature, solar radiation, wind speed, and relative humidity from six different stations operated by California Irrigation Management Information System (CIMIS) located in two different regions of the USA were used in the applications. King-City Oasis Rd., Arroyo Seco, and Salinas North stations are located in San Joaquin region, and San Luis Obispo, Santa Monica, and Santa Barbara stations are located in the Southern region. In the first part of the study, the ANN and M5Tree models were used for estimating ET 0 of six stations and results were compared with the empirical methods. The ANN and M5Tree models were found to be better than the empirical models. In the second part of the study, the ANN and M5Tree models obtained from one station were tested using the data from the other two stations for each region. ANN models performed better than the CIMIS Penman, Hargreaves, Ritchie, and Turc models in two stations while the M5Tree models generally showed better accuracy than the corresponding empirical models in all stations. In the third part of the study, the ANN and M5Tree models were calibrated using three stations located in San Joaquin region and tested using the data from the other three stations located in the Southern region. Four-input ANN and M5Tree models performed better than the CIMIS Penman in only one station while the two-input ANN models were found to be better than the Hargreaves, Ritchie, and Turc models in two stations.

The present study was carried out using artificial neural network (ANN) model for predicting the sodium hazardness, i.e., sodium adsorption ratio (SAR), percent sodium (%Na) residual, Kelly’s ratio (KR), and residual sodium carbonate (RSC) in the groundwater of the Pratapgarh district of Southern Rajasthan, India. This study focuses on verifying the suitability of water for irrigational purpose, wherein more groundwater decline coupled with water quality problems compared to the other areas are observed. The southern part of the Rajasthan State is more populated as compared to the rest of the parts. The southern part of the Rajasthan is more populated as compared to the rest of the Rajasthan, which leads to the industrialization, urbanization, and evolutionary changes in the agricultural production in the southern region. Therefore, it is necessary to propose innovative methods for analyzing and predicting the water quality (WQ) for agricultural use. The study aims to develop an optimized artificial neural network (ANN) model to predict the sodium hazardness of groundwater for irrigation purposes. The ANN model was developed using ‘nntool’ in MATLAB software. The ANN model was trained and validated for ten years (2010–2020) of water quality data. An L-M 3-layer back propagation technique was adopted in ANN architecture to develop a reliable and accurate model for predicting the suitability of groundwater for irrigation. Furthermore, statistical performance indicators, such as RMSE, IA, R, and MBE, were used to check the consistency of ANN prediction results. The developed ANN model, i.e., ANN4 (3-12-1), ANN4 (4-15-1), ANN1 (4-5-1), and ANN4 (3-12-1), were found best suited for SAR, %Na, RSC, and KR water quality indicators for the Pratapgarh district. The performance analysis of the developed model (3-12-1) led to a correlation coefficient = 1, IA = 1, RMS = 0.14, and MBE = 0.0050. Hence, the proposed model provides a satisfactory match to the empirically generated datasets in the observed wells. This development of water quality modeling using an ANN model may help to useful for the planning of sustainable management and groundwater resources with crop suitability plans as per water quality.

BackgroundThis study aimed to develop an artificial neural network (ANN) model for predicting synchronous organ-specific metastasis in lung cancer (LC) patients.MethodsA total of 62,151 patients who diagnosed as LC without data missing between 2010 and 2015 were identified from Surveillance, Epidemiology, and End Results (SEER) program. The ANN model was trained and tested on an 75/25 split of the dataset. The receiver operating characteristic (ROC) curves, area under the curve (AUC) and sensitivity were used to evaluate and compare the ANN model with the random forest model.ResultsFor distant metastasis in the whole cohort, the ANN model had metrics AUC = 0.759, accuracy = 0.669, sensitivity = 0.906, and specificity = 0.613, which was better than the random forest model. For organ-specific metastasis in the cohort with distant metastasis, the sensitivity in bone metastasis, brain metastasis and liver metastasis were 0.913, 0.906 and 0.925, respectively. The most important variable was separate tumor nodules with 100% importance. The second important variable was visceral pleural invasion for distant metastasis, while histology for organ-specific metastasis.ConclusionsOur study developed a “two-step” ANN model for predicting synchronous organ-specific metastasis in LC patients. This ANN model may provide clinicians with more personalized clinical decisions, contribute to rationalize metastasis screening, and reduce the burden on patients and the health care system.

Honey is a naturally sweet and viscous product for which the addition of any substance is prohibited by international regulation. Detection of adulteration in honey is a technical problem: adulteration of honey with invert sugar and syrup may not be reliably detected by direct sugar analysis because its constituents are identical to the major natural components of the honey. Therefore, it is important to develop a rapid and reliable analytical method to detect such additions. We used near-infrared spectroscopy (NIR) combined with principle component analysis (PCA) and artificial neural networks (ANN) modelling to discriminate between honey and corn syrup in adulterated honey. Fifteen honey samples from north-west Croatia (Krapina-Zagorje County) were intentionally supplemented with differing proportions of corn syrup ranging from 10-90%. We collected a total of 460 NIR spectra using the Control Development NIR128L-1.7 spectrophotometer (Control Development, South Bend, Indiana, USA) with their software Spec32 software anda HL-2000 halogen light source. For each of the prepared samples, we measured water content by refractometer (Brouwland, Belgium), conductivity byconductometer (SevenCompact, MettlerToledo, Switzerland), and colour using a PCE-CSM3 colorimeter (PCE Instruments, Germany). Prior to ANN modelling, PCA was used to identify patterns and highlight similarities and differences in data of the individual set of the experiment. The goal of PCA is to extract important information from the data table and to express this information as a set of new orthogonal variables called principal components or factors (PCs or Fs). We conducted PCA of raw spectra using the Unscrambler® X 10.4 software (CAMO software, Norway). Data were divided into ANN model training, test, and validation datasets at a 70:15:15 ratio using the first five PCs. ANNs were calibrated using model training data, and evaluated using model test and model validation datasets for their ability to predict: i) the amount of added adultering substance in honey, ii) water content, iii) conductivity and iv) colour of the adulterated honey. Multiple layer perception (MLP) networks were developed in Statistica v.10.0 software (StatSoft, Tulsa, USA). Back error propagation algorithm available in Statistica v.10.0 was applied for the model training. Model performance was evaluated using R2 and root mean squared error (RMSE) values for model training, test, and validation datasets. Results show that network MLP 5-8-6 with five neurons in the input layer, 8 neurons in the hidden layer and 6 neurons in the output layer predicts the analysed output variables with high precision (R2validation,concentration = 0.995, R2validation,water content = 0.993, R2validation,conductivity = 0.992, R2validation,L = 0.939, R2validation,a = 0.895, R2validation,b = 0.924).

Monthly river flow forecasting using artificial neural network and support vector regression models coupled with wavelet transform

In this study, a feed-forward backpropagation (FFBP) artificial neural network (ANN) model based on multilayer perceptron (MLP) for computing the pull-in voltage value of fixed–fixed micro-actuators is presented. The proposed ANN model is trained using the numerous pull-in voltage value results of clamped–clamped actuators, which have various physical properties, simulated by a software that employs the finite element method (FEM). As the presented ANN model has been trained for the simulation data, it implicitly takes into account the fringing field, mid-plane stretching and size effects. The accuracy and robustness of the results obtained by the proposed ANN model are verified by comparing with those of the theoretical ones through the simulation and experimental studies previously presented in the literature. The average percentage errors for training and testing data are found to be 0.30 and 0.40 %, respectively, while the maximum percentage error for literature data is calculated as 1.96 %. These results show that the main advantage of the presented ANN model gives satisfying pull-in voltage results over solution space of the simulations with effortless way.

Purpose – This paper aims to predict the needle penetration force (NPF) in denim fabrics using the artificial neural network (ANN) and multiple linear regression (MLR) models based on the effects of various sewing parameters. Design/methodology/approach – In order to design the ANN and MLR models, four parameters including fabric weight, number of fabric layers, weave pattern, and sewing needle size are taken into account as the input parameters and NPF as the output parameter. According to these parameters, 140 samples of data were resulted. Each sample was tested five times. From these 140 data (input-output data pairs), 112 were used for training the ANN and MLR models and 28 samples were used to test the performance of ANN and MLR. Also, the NPF was measured on the Instron tensile tester to simulate sewing process. Findings – The results indicated that the NPF in denim fabrics can be well predicted in terms of sewing parameters by using ANN and MLR models, in which the ANN model exhibits greater performance than MLR (RANN=0.989> RMLR=0.901). Research limitations/implications – The NPF measurement method is limited at low speed. Originality/value – Using the ANN model for forecasting NPF in denim fabrics can help the garment manufactures to produce high-quality denim products and improve the sewing process through reducing seam damage. The NPF could be also measured in the cycle loading conditions similar to sewing machine process by using a special designed tools mounted on the Instron tensile tester.

ABSTRACTStatistical methods such as Poisson distribution, negative binomial regression (NB), and zero inflated negative binomial regression (ZINB) have generally been used in road safety studies to establish the complex relationships between variables. Over the last few years, the artificial neural networks (ANN) model has also been used. The ANN model does not have any prior limitations such as the equality condition of mean and variance observed in Poisson regression. However, though the ANN model has been used in the analysis of different accident types, to the best of our knowledge, no study has used the ANN model for establishing this relationship with truck accident data on divided multilane interurban roads. In this study, the road sections D750/07–D750/15 in Ankara–Aksaray–Eregli, Turkey, were considered and truck accident data from 2008 to 2011 were analyzed using NB and ANN. The analysis show that the ANN model has lower errors and higher R2 values than NB and performs slightly better than NB for predicting the number of trucks involved in accidents. Based on a comparison of performances the study concludes, that ANN could be used as an alternative model for analyzing truck accident data on divided multilane interurban roads in Turkey.

The power capacity of solar photovoltaics (PVs) in Korea has grown dramatically in recent years, and an accurate estimation of solar resources is crucial for the efficient management of these solar PV systems. Since the number of solar irradiance measurement sites is insufficient for Korea, satellite images can be useful sources for estimating solar irradiance over a wide area of Korea. In this study, an artificial neural network (ANN) model was constructed to calculate hourly global horizontal solar irradiance (GHI) from Korea Communication, Ocean and Meteorological Satellite (COMS) Meteorological Imager (MI) images. Solar position variables and five COMS MI channels were used as inputs for the ANN model. The basic ANN model was determined to have a window size of five for the input satellite images and two hidden layers, with 30 nodes on each hidden layer. After these ANN parameters were determined, the temporal and spatial applicability of the ANN model for solar irradiance mapping was validated. The final ANN ensemble model, which calculated the hourly GHI from 10 independent ANN models, exhibited a correlation coefficient (R) of 0.975 and root mean square error (RMSE) of 54.44 W/m² (12.93%), which were better results than for other remote-sensing based works for Korea. Finally, GHI maps for Korea were generated using the final ANN ensemble model. This COMS-based ANN model can contribute to the efficient estimation of solar resources and the improvement of the operational efficiency of solar PV systems for Korea.

In this research work, we developed a mathematical model of a digital currency market, involving daily closing price as a function of time. We proposed the Artificial Neural Network (ANN) model. We observed that our ANN model was able to predict the daily closing price of Bitcoin and also make six weeks forecast to a reasonable degree of accuracy. We equally observe that the time dependent ANN model can actually give digital currency traders and investors a clue on when to trade off their digital assets with minimum risk. We therefore, recommend that ANN model should be incorporated into digital currency trading platforms as a signal tool to enable digital currency traders take more informed and less risky trading decisions. From our findings, we would advise traders who wish to employ ANN model to consider a smaller time frame say a few weeks’ time interval for their predictions. We observed also that ANN models have limitations when it comes to manual computation or implementation in Microsoft Excel, especially when dealing with very large input values. This is because of the saturation characteristic of our ANN inner layer activation function (viz; tanh function) which can lead to identical output values for different input values, making it difficult to replicate the ANN model's behavior. Furthermore, ANN models often involve complex interactions between multiple neurons, layers, and activation functions, which can be challenging to replicate manually.

According to the thought of intelligent forecasting and hybrid forecasting, an Intelligent Hybrid (IH) model for short-term traffic flow forecasting was presented. The IH model had three sub-models: History Mean (HM) model, Artificial Neural Network (ANN) model and the Fuzzy Combination (FC) model. By means of the good static stabilization character of HM method, the HM model predicted the traffic flow by the Single Exponential Smoothing method based on the historical traffic data. Otherwise, the ANN model was a 1.5-layer feed-forward neural network built by some common S-function neurons. Because of the strong dynamic nonlinear mapping ability of ANN, the ANN model can estimate the actual traffic flow in a very precise and satisfactory sense. The FC model mixed the two individual forecasting results by fuzzy logic and its output was regarded as the final forecasting of the traffic flow. Factual application results show that the IH model, which takes advantage of the unique strength of the HM model and the ANN model, can produce more precise forecasting than that of two individual models. Thus, the IH model can be an efficient method to the short-term traffic flow forecasting.

Nitric acid production plants emit small amounts of nitrogen oxides (NOx) to the environment. As the regulatory authorities demand the reduction of the resulting air pollution, existing plants are looking for economical ways to comply with this demand. Several Artificial Neural Networks (ANN) models were trained from several months of operating plant data to predict the NOx concentration in the tail gas, and their total amount emitted the environment. The training of the ANN model was done by the Guterman-Boger algorithm set that generates a non-random initial connection weights, suggests a small number of hidden neurons, avoids, and escapes from, local minima encountered during the training. The ANN models gave small errors, 0.6% relative error on the NOx concentration prediction and 0.006 kg/hour on daily emission in the 20-45 kg NOx/hour range. Knowledge extraction from the trained ANN models revealed the underlying relationships between the plant operating variables and the NOx emission rate, especially the beneficial effect of cooling the absorbed gas and reticulating liquids in the absorption towers. Clustering the data by the patterns of the hidden neurons outputs of auto-associative ANN models of the same data revealed interesting insights.

The purpose of this study was to validate the use of artificial neural network (ANN) models for predicting quality of life (QOL) after breast cancer surgery and to compare the predictive capability of ANNs with that of linear regression (LR) models. The European Organization for Research and Treatment of Cancer Quality of Life Questionnaire and its supplementary breast cancer measure were completed by 402 breast cancer patients at baseline and at 2 years postoperatively. The accuracy of the system models were evaluated in terms of mean square error (MSE) and mean absolute percentage error (MAPE). A global sensitivity analysis was also performed to assess the relative significance of input parameters in the system model and to rank the variables in order of importance. Compared to the LR model, the ANN model generally had smaller MSE and MAPE values in both the training and testing datasets. Most ANN models had MAPE values ranging from 4.70 to 19.96 %, and most had high prediction accuracy. The ANN model also outperformed the LR model in terms of prediction accuracy. According to global sensitivity analysis, pre-operative functional status was the best predictor of QOL after surgery. Compared with the conventional LR model, the ANN model in the study was more accurate for predicting patient-reported QOL and had higher overall performance indices. Further refinements are expected to obtain sufficient performance improvements for its routine use in clinical practice as an adjunctive decision-making tool.

Multivariate regression (MVR) and different artificial neural network (ANN) models developed for optical transparency of conductive polymer nanocomposite films