Search

This research aims to determine the legal awareness of juveniles in archipelagic areas using social media. This type of research is empirical legal research. The data was collected using a questionnaire and direct interviews. The data obtained in this research were then analyzed descriptively qualitatively with a sociological juridical approach to conclude the research's object. The results show that most of the legal awareness of juveniles in archipelagic areas is still low using social media. In this case, most juveniles do not know (54%) and do not understand (66%) on regulations for using social media. In addition, most behaviors of juveniles have violated (72%) of the regulations for using social media. Therefore, it is recommended for parents, teachers, and law enforcement officers to provide assistance and education to juveniles in using social media. In addition, law enforcement and local government officials can involve the participation of the community to be more intense in conducting legal counseling for juveniles. In this case, juveniles in archipelagic areas will have more legal awareness and comply with laws and regulations in the future.

Vehicle suspension system would complement vehicle safety in terms of ride comfort, handling and braking, and also isolation from road irregularities. This can be better achieved by introducing a semi-active suspension system where the magneto-rheological (MR) fluid is used to manipulate its damping characteristics. This paper explains the controller development for MR semi-active suspension system. The process started by developing the non-parametric linearized data driven (NPLDD) double input with proportional-integral (PI) controller as an inner loop algorithm for MR damper modelling and force tracking control; then, Improved PI controller as an outer loop control algorithm for vehicle response improvement, and finally validation work using quarter-car test bench. All the algorithm development is based on Perusahaan Otomobil Nasional (PROTON) vehicle characteristics and specifications. From the simulation results, it is verified that the NPLDD double input model together with the PI control strategy has the capability to track the desired damping force well. Meanwhile, the Improved PI controller for the outer loop is capable to reduce the magnitude of vehicle body acceleration and displacement. The validation at the quarter-car test bench also shows significant reduction of these two parameters. As a result, the PI controller for the MR semi-active suspension system is found sufficient for vehicle safety improvement.

Adaptive Cruise Control (ACC) is introduced to allow the subject vehicle to follow a target vehicle at a pre-selected time gap, up to a driver-selected velocity by controlling the engine and/or service brakes as defined in SAE J2399. A typical ACC ensures passenger comfort by minimizing the subject vehicle’s rate of acceleration and deceleration when following a target vehicle. However, if the ACC rate of velocity change is too slow, the subject vehicle might not be able to follow the target vehicle at the set time-gap. On the other hand, if the ACC responds too quickly to the changes in the target vehicle velocity, it might cause passenger discomfort and reduce energy efficiency due to the high rate of acceleration and deceleration. Thus, a research gap in ACC is identified, and this paper seeks to investigate an optimal ACC for driving comfort and fuel economy. This paper models the performance of a passenger vehicle with ACC using feedback control in Matlab Simulink. Three types of ACC control systems are simulated and compared; proportional-integral (PI) control, fuzzy logic control (FLC), and Linear-Quadratic Regulator (LQR). The performance of the control systems is simulated using set test scenarios, and an optimal ACC design is selected based on minimum acceleration, deceleration, and jerk performance, as defined in ISO 15622, following performance and minimum fuel consumption. Results show that the LQR controller has the best performance in meeting the ISO 15622, the lowest target vehicle the following root mean square error (RMSE), and the second-lowest New European Drive Cycle (NEDC) urban cycle fuel consumption.

The Sumatran Rhinoceros Dicerorhinus sumatrensis is on the edge of extinction. The decline of this species was initially attributed to poaching and habitat loss, but evidence presented here indicates that reproductive failure has also been a significant cause of loss, and continues to affect wild populations. Indonesia’s remaining populations of Sumatran Rhino are small and scattered, with limited access to breeding opportunities with unrelated mates. This leaves them subject to inbreeding and isolation-induced infertility, linked to fertility problems analyzed here. Sumatran Rhino females in captivity showed high rates (>70%) of reproductive pathology and/or problems with conception, which has significantly hindered the breeding program. Technological advances enabling examination immediately after capture revealed similarly high rates and types of reproductive problems in individuals from wild populations. The last seven Sumatran Rhino females captured were from areas with small declining populations, and six had reproductive problems. Going forward, capturing similarly compromised animals will take up valuable space and resources needed for fertile animals. The high risk of infertility and difficulty of treating underlying conditions, coupled with the decreasing number of remaining animals, means that the success of efforts to build a viable captive population will depend upon utilizing fertile animals and applying assisted reproductive techniques. Decades of exhaustive in situ surveys have not provided information relevant to population management or to ascertaining the fertility status of individual animals. Thus the first priority should be the capture of individuals as new founders from areas with the highest likelihood of containing fertile rhinos, indicated by recent camera trap photos of mothers with offspring. In Sumatra these areas include Way Kambas and parts of the Leuser ecosystem.

Recent development of Advance Driver Assistance System (ADAS) has seen various advancement in object detection for vehicle vision system, particularly on the detection of other vehicles, pedestrians, road lane and signage. While these detections can provide assistant to avoid road accidents, they still lack to include road condition factors that also contributed to road accidents in Malaysia. This paper proposes a detection of the road peculiarities such as pothole and road bumps to act as additional safety feature in ADAS. With the breakthrough of deep learning in solving image recognition problems, this work takes advantage of Single Shot Detector (SSD)-MobileNetV2 as the detection algorithm, implemented on the real-time. The training images for potholes and road bumps taken from the Malaysia roads are fed into the detection model, and then the pre-trained weights are fine-tuned over the training process. The results show that the detection algorithm can predicts the potholes and road bumps, while exhibit the detection accuracy and confidence limitation due to the variety of shape and pattern of potholes and road bumps. Testing the detection algorithm with NVIDIA Jetson Nano yielded about 20 frames per second (fps), suitable for real-time applications.

In this work, the plastic crystal polymer electrolytes (PCPEs), composed of polyacrylonitrile (PAN), succinonitrile (SN) and lithium bis(trifluoromethane)sulfonimide (LiTFSI) were prepared. The concentrations of lithium salt were varied by weight percentage from 10 wt% to 50 wt%. The ionic conductivity of the PCPE films increases with the increase of lithium salt, where the highest value recorded is in the order of ~10−2 S cm−1. The temperature-dependence conductivity analysis shows that the PCPE films exhibit Arrhenius behaviour when subjected to the temperature range from 303 K to 343 K. The decrease in crystallinity was confirmed by X-ray diffraction (XRD) and Differential Scanning Calorimetry (DSC) analyses. The cationic transport number also increases with the increase of salt which corresponds well to their conductivity values. It is found that the films are electrochemically stable up to ~3.6 V as revealed by the linear sweep voltammetry (LSV) analysis. The cyclic voltammetry (CV) plots of the films shown no substantial change in the redox peaks which mean that the charge transfer reaction is reversible.

A new gel polymer electrolytes composed of poly(vinylidene chloride-co-acrylonitrile) (PVdC-co-AN) co-polymer and plastic crystal succinonitrile (SN) were prepared to form plastic crystal gel polymer electrolytes (PGPEs) with the variation of magnesium triflate (MgTf2) salt from 5 to 30 wt.%. The room temperature and temperature dependence studies of the PGPE films were determined by using AC impedance spectroscopy. The conductivity versus reciprocal temperature plot has shown a nonlinear behavior. The ionic and cationic transport numbers were evaluated by DC polarization method and the combination of DC polarization and AC impedance methods, respectively to determine the charge carrier species in the PGPE films. To study the interactions among the constituents in the PGPEs as well as to confirm the complexation between them, Fourier transform infrared spectroscopy (FTIR) was carried out. The analysis of FTIR spectra was further investigated by deconvolution of the FTIR spectra to prove the dependability of ionic conductivity and transport number with the presence of free ions, ion pairs, and ion aggregates in the PGPEs.

The availability of a wetland environment-based practicum module as a form of local wisdom in South Kalimantan has not been widely developed, even though the development of a wetland-based practicum module is needed to create more meaningful learning. This study aimed to describe the validity of the fluid physics practicum module based on the wetland environment. The research method used was R&D with a 4D model. The research instrument used was in the form of a validation sheet, assessed by three validators. The results showed that the validity of the practicum module was very valid. This practicum module that was declared as valid can be used in the learning process. At the next stage, the practicality and effectiveness of the fluid physics practicum module based on the wetland environment can be measured.

This study discussed about of prototype of electrical generator device from piezoelectric with serial configuration - based on mechanical pressure. A prototype of a piezoelectric-based power plant that is able to convert mechanical energy (footing) or object load on piezoelectric into electrical energy was the aim of this research. The piezolectric generation system can be applied to serve limited loads and can be a support system for a distributed generation system. The method of research was device design and experimental testing. The experimental testing was carried out by varying pressures on piezoelectric to produce the desired output voltage. In testing, the pressure was generated by two different objects, a human footing and a object load with varying weight. The average voltage produced on human footing was 9.736 V for 60 kg body weight. The average voltage produced on object loads was 1.33 V for 5 kg object weight. This research has succeeded in making a prototype of a power plant using piezoelectric.

Morinda citrifolia (MC) is a herb that is famous for its various medicinal properties. However, there has been little research on its effect on increased stamina and physical fatigue reduction. The purpose of this study was to determine the effect of high-dose oral intake of 50% ethanol MC leaf extract on exercise performance and physical fatigue in an animal model. Thirty female Sprague-Dawley rats were divided into five groups; Control (C); Sedentary (S); Exercise (EX); MC leaf extract 1000 mg/kg/day (MC); MC leaf extract 1000 mg/kg/day with Exercise (MC+EX). The endurance test showed a significant difference in endurance time between EX and MC groups, whereby intake of MC leaf extract without daily exercise lowered endurance level by 68.75%. These findings correlated with serum fatigue indicator analysis whereby significant increases in CK levels of MC group, and LDH levels in MC and MC+EX groups were noted when compared to other groups. Electron microscopy analysis on cardiac muscle tissue showed ultrastructural injuries in muscle fibers of the EX, MC and MC+EX groups whereas injuries to the gastrocnemius myofibres were not so prominent. Based on the study findings, it can be concluded that the consumption of 1000 mg/kg/day 50% ethanol MC extract disturbed internal physiology and resulted in injury to the muscle tissues thus lowering rats endurance capability. This shows that a high dose has unfavourable effects on an animal model while a lower dose may be more appropriate.

Vegetable is an essential daily diet item for the people of Malaysia. This work addressed the radiation and heavy metal exposure scenarios through the consumption of vegetables. Kuala Selangor is located in Sungai Selangor estuary in the west coast of Peninsular Malaysia, which is susceptible to pollution load due to the presence of large-scale industrial and human activities. Radioactivity and heavy metals level in human diet is of particular concern for the assessment of possible radiological and chemical hazards to human health. Therefore, a comprehensive study was carried out to determine the radioactivity levels ((226)Ra, (228)Ra and (40)K) and heavy metal concentrations (Cr, As, Cd, Mn, Mg, Al, Sr, Rb, Sb, Ba, Hg, Fe, Ni, Zn, Cu, Bi and Pb) in 10 varieties of vegetable collected from different farmlands in Kuala Selangor region. The committed doses for (226)Ra, (228)Ra and (40)K due to consumption of vegetables were found 16.6±1.3, 23.6±1.7 and 58±5 µSv y(-1), respectively, with a total of 98±8 µSv y(-1). This dose imposes no significant threat to human health. The estimated cancer risk shows that probability of increase in cancer risk from daily intake of vegetables is only a minor fraction of International Commission on Radiological Protection values. The concentrations of heavy metal were below the daily intake recommended by the international organisations.

The purpose of this study is to determine the factors that influence the relational satisfaction between chicken retailers and chicken suppliers in Terengganu. In this study, relational satisfaction as the dependent variable and six independent variables, namely trust, commitment, service quality, personal, organization and the business environment. The independent variable in this research is divided into two components which are labelled as intangible factors and tangible factors. The result discovered that the correlation analysis, all six independent variables are positively and significantly correlated to the dependent variable. The regression analysis revealed that only the intangible variables have a significant and positive relationship with the dependent variable. Moreover, the trust factor is the most influencing factor among the retailers followed by service quality and commitment. Keywords: relational satisfaction; trust; service quality; chicken retailers; suppliers

Polyvinylidenefluoride-co-hexafluoropropylene (PVdF-HFP) gel polymer electrolytes (GPEs) consisting of sodium trifluoromethane-sulfonate or sodium triflate were prepared for sodium ion battery application. The GPEs at different weight percent ratios of salt content were prepared by solution casting technique. The effects of changes in salt concentration on the structural and electrical properties were investigated. FTIR studies confirm that conductivity enhancement is due to the increase in free ions and decrease in ion aggregates. Impedance studies indicate that the highest ionic conductivity of 2.50 × 10-3 S / cm was observed with the addition of 25 wt. % of sodium salt concentration in GPEs system obtained at ambient temperature. The transport number of sodium ions in the GPEs was evaluated using the combination of AC impedance spectroscopy and DC polarization techniques.

Teaching materials that link aspects of science, environment, technology, and society, as well as teaching materials that contain character values, are not yet in school. For this reason, this study is conducted to describe the feasibility of teaching materials developed in terms of the aspects of validity, practicality, effectiveness, and character achievement. This research includes research and development (R&D) that refers to the ADDIE model. The subjects of the trial from this study were 33 VIIIH class students of SMP 1 Banjarmasin. The research instruments used included the validation sheet, lesson plan implementation sheet, learning outcome test, and self-assessment sheet. The results of data analysis show that the natural science teaching material developed has a very good validation category; practical, which can be done very well; effectiveness with effective categories; and the achievement of students’ character is a very good category. It was concluded that natural science teaching materials character-based appropriate were used on the subject matter of vibrations, waves, and sounds for middle school in science technology and society (STS) approach.

Implementation of rehabilitation for prisoners in an effort to return prisoners to a good society is very important to do, not only material or spiritual, but both must be balanced. The problem in this study was how to implement a prisoner rehabilitation system so that inmates were aware of mistakes, improved themselves and did not repeat criminal acts in correctional institution Class IIA Palopo. This type of research is empirical legal research that is descriptive qualitative. The implementation of rehabilitation in the correctional system affects the prisoners who undergo their criminal periods while in prison. Guidance of prisoners in Prison Class IIA Palopo is carried out through personality development which includes (religious formation, legal awareness development), and self-development including prisoner skills training and assimilation processes. The research recommendation is that penitentiary Officers intensify the formation of personality and independence of prisoners with the aim of no longer committing crimes and can socialize themselves inmates in the community. In addition, it is necessary to collaborate with the Ministry of Religion to intensify the religious rehabilitation of prisoners in the context of socialization in the community.Keywords: Penitentiary, Prisoners, and Rehabilitation.

In this work, we have prepared systems of poly(vinylidene chloride-co-acrylonitrile) (PVdC-co-AN) based gel polymer electrolytes (GPEs) which are single plasticized-GPEs and double plasticized-GPEs. Both systems comprised plastic crystal succinonitrile SN to form plastic crystal gel polymer electrolyte (PGPE) films. The ionic conductivity of the PGPE films were analysed by means of a.c. impedance spectroscopy at room temperature as well as at the temperature range of 303 K to 353 K. The temperature dependence ionic conductivity was found to obey the VTF rule. To study the interactions among the constituents in the PGPEs, Fourier Transform Infrared Spectroscopy (FTIR) was carried out and hence, the complexation between them has also been confirmed.

Silk fibroin (SF), isolated from silkworm (Bombyx mori) cocoons, is a natural biodegradable polymer. Over the past decade, there was some interest in using SF as a biomedical material. As part of a project to develop tissue-engineered constructs for the surgical restoration of the ocular surface (cornea, conjunctiva), we have investigated the capacity of SF to function as a substratum for the attachment and growth of corneal stem/progenitor cells harvested from the corneoscleral limbus of donor human corneal tissue. SF membranes were produced from cocoons following a protocol involving successive dissolution steps, filtration, dialysis, evaporation, and methanol treatment. Human limbal epithelial cells were harvested from donor tissue and seeded onto SF membranes. After 5 days, the culture was fixed and stained with specific agents to visualize the cells. The study indicated profuse cellular attachment and growth. SF membranes appear to be suitable as a substratum for the repair of damaged ocular surface.

This research examines the practice of green property management (GPM) for commercial buildings. Property management is the process of taking care of properties so as to meet the interests of all people having an interest in the property. The extent of damage to the environment from current unsustainable living today has driven building users and operators to adopt a more environmental practice towards property asset management. Nevertheless, practice for GPM experiences different understanding on the concept and hence, there is no standard practice towards it. In this research, the procedures and methods of developing the green property management framework examines some of the property management requirements for the practice of GPM, in particular, for commercial buildings. In particular, it examines the work procedures in current property management practice by property professionals and to integrate performance measurement by the GBI in designing the framework for GPM practice.

A 7-year-old boy, referred with lymphoma, presented with prolonged fever and intra-abdominal lymphadenopathy demonstrated on computed tomography (CT) of the abdomen. Blood culture isolated Penicillium marneffei. The patient was subsequently proven serologically to be positive for human immunodeficiency virus (HIV). Treatment with amphotericin B followed by itraconazole was successful. A high level of clinical suspicion and awareness is necessary for early diagnosis of penicilliosis, especially in an era of an increasing prevalence of HIV in this region.

The mtDNA and Y chromosomes of the Malay population of peninsular Malaysia show high levels of diversity. Both the Malay and the indigenous Orang Asli populations showed a similar level of divergence between the different Y chromosome haplotypes.

The purpose of this study is to analyze the important of relationship marketing in the hotel industry performance regarding Permai Hotel that represents a three-star hotel. The research involved 108 employees of Permai Hotel from a variety of departments. A simple random sampling was used and the data are analyzed using reliability analysis, descriptive frequency analysis, Pearson’s correlation analysis, and regression analysis. The result shows that there is a moderate and positive relationship between communication, shared value, empathy and reciprocity with organizational performances. Although the association between the variables is moderately positive, it indicates that the independent variable influences the organization’s performance is significant. It is also discovered that the most influential components of relationship marketing orientation that affect organizational performance is reciprocity. Keywords: relationship marketing; marketing orientation; organizational performance; hotel industry; trust

Various types of internal antennas suitable for digital TV receiver have been investigated. These include rectangular plate loop structure, 1/2λ monopole, and L-dipole. Evaluations of the internal antennas were done, numerically and experimentally, to obtain some electrical properties such as S 11 , VSWR, gain and radiation pattern. All the fabricated antennas are incorporated inside the TV receiver body. This allows more flexible, moveable, compact, and inexpensive construction of digital TV in the future.

This paper proposes non-Gaussian models for parametric spectral estimation with application to event-related desynchronization (ERD) estimation of nonstationary EEG. Existing approaches for time-varying spectral estimation use time-varying autoregressive (TVAR) state-space models with Gaussian state noise. The parameter estimation is solved by a conventional Kalman filtering. This study uses non-Gaussian state noise to model autoregressive (AR) parameter variation with estimation by a Monte Carlo particle filter (PF). Use of non-Gaussian noise such as heavy-tailed distribution is motivated by its ability to track abrupt and smooth AR parameter changes, which are inadequately modeled by Gaussian models. Thus, more accurate spectral estimates and better ERD tracking can be obtained. This study further proposes a non-Gaussian state space formulation of time-varying autoregressive moving average (TVARMA) models to improve the spectral estimation. Simulation on TVAR process with abrupt parameter variation shows superior tracking performance of non-Gaussian models. Evaluation on motor-imagery EEG data shows that the non-Gaussian models provide more accurate detection of abrupt changes in alpha rhythm ERD. Among the proposed non-Gaussian models, TVARMA shows better spectral representations while maintaining reasonable good ERD tracking performance.

Catalytic reactions of n-octane on sequential cambinatians of HY and HZSM-5, as well as on random mixtures of the zeolites, have been examined. Theoretical product distributions have been calculated for reactions on these catalyst combinations by assuming relative contributions based on the individual product distributions as well as the conversions on the individual components. Effects due to interactions of products of one catalyst with the other in a binary combination can be identified by comparing theoretical and observed product distributions. Predicted values from this approach provide a good approximation when considering the product distributions in terms of total hydrocarbons formed at each carbon number. Significant differences are found between observed and predicted values in the relative proportion of paraffins and olefins at a particular carbon number, and also in the ratios of branched to linear paraffin isomers. Reaction processes attributable for these deviations have been identified as hydrogen transfer processes, particularly on HY, and paraffin isomerization on HZSM-5.

Catalytic reactions of n-octane on sequential cambinatians of HY and HZSM-5, as well as on random mixtures of the zeolites, have been examined. Theoretical product distributions have been calculated for reactions on these catalyst combinations by assuming relative contributions based on the individual product distributions as well as the conversions on the individual components. Effects due to interactions of products of one catalyst with the other in a binary combination can be identified by comparing theoretical and observed product distributions. Predicted values from this approach provide a good approximation when considering the product distributions in terms of total hydrocarbons formed at each carbon number. Significant differences are found between observed and predicted values in the relative proportion of paraffins and olefins at a particular carbon number, and also in the ratios of branched to linear paraffin isomers. Reaction processes attributable for these deviations have been identified as hydrogen transfer processes, particularly on HY, and paraffin isomerization on HZSM-5.

This paper reports work done to improve the modeling of complex processes when only small experimental data sets are available. Neural networks are used to capture the nonlinear underlying phenomena contained in the data set and to partly eliminate the burden of having to specify completely the structure of the model. Two different types of neural networks were used for the application of pulping problem. A three layer feed forward neural networks, using the Preconditioned Conjugate Gradient (PCG) methods were used in this investigation. Preconditioning is a method to improve convergence by lowering the condition number and increasing the eigenvalues clustering. The idea is to solve the modified odified problem M-1 Ax= M-1b where M is a positive-definite preconditioner that is closely related to A. We mainly focused on Preconditioned Conjugate Gradient- based training methods which originated from optimization theory, namely Preconditioned Conjugate Gradient with Fletcher-Reeves Update (PCGF), Preconditioned Conjugate Gradient with Polak-Ribiere Update (PCGP) and Preconditioned Conjugate Gradient with Powell-Beale Restarts (PCGB). The behavior of the PCG methods in the simulations proved to be robust against phenomenon such as oscillations due to large step size.

Since the presentation of the backpropagation algorithm, a vast variety of improvements of the technique for training a feed forward neural networks have been proposed. This article focuses on two classes of acceleration techniques, one is known as Local Adaptive Techniques that are based on weightspecific only, such as the temporal behavior of the partial derivative of the current weight. The other, known as Dynamic Adaptation Methods, which dynamically adapts the momentum factors, α, and learning rate, η, with respect to the iteration number or gradient. Some of most popular learning algorithms are described. These techniques have been implemented and tested on several problems and measured in terms of gradient and error function evaluation, and percentage of success. Numerical evidence shows that these techniques improve the convergence of the Backpropagation algorithm., {"references":["Magoulas, G.D., Plagianakos, V.P., & Vrahatis, M.N. Globally\nConvergent Algorithms with Local Learning Rates. In IEEE\nTransactions on Neural Networks, Vol 13, No 3, 774 - 779, 2002.","Silva F.M. & Almeida L.B., Acceleration techniques for the\nbackpropagation algorithm. Neural Networks EURASIP Workshop,\nSesim, 1990.","Jacobs R. A., Increased rates of convergence through learning rate\nadaptation, Neural Networks, 1(4), 295-308, 1988.","Tollenaere, T., SuperSab:fast adaptive backpropagation with good\nscaling properties. Neural Networks, 3(5), 1990.","Fahlman, S.E., An empirical study of learning speed in backpropagation\nnetworks, Technical Report, CMU-CS-88-162, 1988","Vrahatis, M.N., Magoulas, G.D. & Plagaianakos, V.P. Convergence\nAnalysis of the Quickprop Method. University of Patras and Athens,\n1999.","Riedmiller, M. & Braun, H., A direct adaptive method for faster\nbackpropagation learning. The RPROP algorithm. In Proceedings of the\nIEEE International Conference on Neural Networks (ICNN) (Ruspini,\nH. ed), p. 586-591, 1993, San Francisco.","Evans, D.J. and Zainuddin, Z., Acceleration of the backpropagation\nthrough dynamic adaptation of the momentum. Neural, Parallel &\nScientific Computations, 5(3), 1997, 297-308. (see also Internal Report\nNo.1028, PARC, Loughborough University of Tech., U.K., 1996).","Zainuddin Z. & Evans D.J., Acceleration of the backpropagation through\ndynamic adaptation of the learning rate, Int. Journal of Computer\nMathematics, 334, 1-17, 1997 (see also Internal report No. 1029, PARC,\nLoughborough University of Tech., U.K. 1996).\n[10] Zainuddin Z. & Sathasivam S. , Modeling nonlinear relationships in\necology and biology using neural networks, Proceedings of the National\nWorkshop in Ecological and environmental modeling (ECOMOD), Sept.\n3-4, 2001.\n[11] Zainuddin Z. & Ahmad Fadzil, M.H., Training Feedforward Neural\nNetworks: An Algorithm Giving Improved Convergence. Proceedings of\nresearch & Development in Computer Science & its Application,p.98-\n102, 1998, Penang, Malaysia."]}

This paper reports work done to improve the modeling of complex processes when only small experimental data sets are available. Neural networks are used to capture the nonlinear underlying phenomena contained in the data set and to partly eliminate the burden of having to specify completely the structure of the model. Two different types of neural networks were used for the application of Pulping of Sugar Maple problem. A three layer feed forward neural networks, using the Preconditioned Conjugate Gradient (PCG) methods were used in this investigation. Preconditioning is a method to improve convergence by lowering the condition number and increasing the eigenvalues clustering. The idea is to solve the modified problem where M is a positive-definite preconditioner that is closely related to A. We mainly focused on Preconditioned Conjugate Gradient- based training methods which originated from optimization theory, namely Preconditioned Conjugate Gradient with Fletcher-Reeves Update (PCGF), Preconditioned Conjugate Gradient with Polak-Ribiere Update (PCGP) and Preconditioned Conjugate Gradient with Powell-Beale Restarts (PCGB). The behavior of the PCG methods in the simulations proved to be robust against phenomenon such as oscillations due to large step size.