Abstract: This paper presents the methodology from machine
learning approaches for short-term rain forecasting system. Decision
Tree, Artificial Neural Network (ANN), and Support Vector Machine
(SVM) were applied to develop classification and prediction models
for rainfall forecasts. The goals of this presentation are to
demonstrate (1) how feature selection can be used to identify the
relationships between rainfall occurrences and other weather
conditions and (2) what models can be developed and deployed for
predicting the accurate rainfall estimates to support the decisions to
launch the cloud seeding operations in the northeastern part of
Thailand. Datasets collected during 2004-2006 from the
Chalermprakiat Royal Rain Making Research Center at Hua Hin,
Prachuap Khiri khan, the Chalermprakiat Royal Rain Making
Research Center at Pimai, Nakhon Ratchasima and Thai
Meteorological Department (TMD). A total of 179 records with 57
features was merged and matched by unique date. There are three
main parts in this work. Firstly, a decision tree induction algorithm
(C4.5) was used to classify the rain status into either rain or no-rain.
The overall accuracy of classification tree achieves 94.41% with the
five-fold cross validation. The C4.5 algorithm was also used to
classify the rain amount into three classes as no-rain (0-0.1 mm.),
few-rain (0.1- 10 mm.), and moderate-rain (>10 mm.) and the overall
accuracy of classification tree achieves 62.57%. Secondly, an ANN
was applied to predict the rainfall amount and the root mean square
error (RMSE) were used to measure the training and testing errors of
the ANN. It is found that the ANN yields a lower RMSE at 0.171 for
daily rainfall estimates, when compared to next-day and next-2-day
estimation. Thirdly, the ANN and SVM techniques were also used to
classify the rain amount into three classes as no-rain, few-rain, and
moderate-rain as above. The results achieved in 68.15% and 69.10%
of overall accuracy of same-day prediction for the ANN and SVM
models, respectively. The obtained results illustrated the comparison
of the predictive power of different methods for rainfall estimation.
Abstract: In this paper, a neural network tuned fuzzy controller
is proposed for controlling Multi-Input Multi-Output (MIMO)
systems. For the convenience of analysis, the structure of MIMO
fuzzy controller is divided into single input single-output (SISO)
controllers for controlling each degree of freedom. Secondly,
according to the characteristics of the system-s dynamics coupling, an
appropriate coupling fuzzy controller is incorporated to improve the
performance. The simulation analysis on a two-level mass–spring
MIMO vibration system is carried out and results show the
effectiveness of the proposed fuzzy controller. The performance
though improved, the computational time and memory used is
comparatively higher, because it has four fuzzy reasoning blocks and
number may increase in case of other MIMO system. Then a fuzzy
neural network is designed from a set of input-output training data to
reduce the computing burden during implementation. This control
strategy can not only simplify the implementation problem of fuzzy
control, but also reduce computational time and consume less
memory.
Abstract: We investigated the effects of modified
preprogrammed training mode Chase Trainer from Balance Trainer
(BT3, HurLab, Tampere, Finland) on athlete who experienced
unilateral Patellofemoral Pain Syndrome (PFPS). Twenty-seven
athletes with mean age= 14.23 ±1.31 years, height = 164.89 ± 7.85
cm, weight = 56.94 ± 9.28 kg were randomly assigned to two groups:
experiment (EG; n = 14) and injured (IG; n = 13). EG performed a
series of Chase Trainer program which required them to shift their
body weight at different directions, speeds and angle of leaning twice
a week for duration of 8 weeks. The static postural control and
perceived pain level measures were taken at baseline, after 6 weeks
and 8 weeks of training. There was no significant difference in any of
tested variables between EG and IG before and after 6-week the
intervention period. However, after 8-week of training, the postural
control (eyes open) and perceived pain level of EG improved
compared to IG (p
Abstract: During the last few decades in the academic field, the
debate has increased on the effects of social geography on the
opportunities of socioeconomic integration. On one hand, it has been
discussed how the contents of the urban structure and social
geography affect not only the way people interact, but also their
chances of social and economic integration. On the other hand, it has
also been discussed how the urban structure is also constrained and
transformed by the action of social actors. Without questioning the
powerful influence of structural factors, related to the logic of the
production system, labor markets, education and training, the
research has shown the role played by place of residence in shaping
individual outcomes such as unemployment. In the context of this
debate the importance of territory of residence with respect to the
problem of unemployment has been highlighted.
Although statistics of unemployment have already demonstrated
the unequal incidence of the phenomenon in social groups, the issue
of uneven territorial impact on the phenomenon at intra-urban level
remains relatively unknown.
The purpose of this article is to show and to interpret the spatial
patterns of unemployment in the city of Porto using GIS (Geographic
Information System - GIS) technology. Under this analysis the
overlap of the spatial patterns of unemployment with the spatial
distribution of social housing, allows the discussion of the
relationship that occurs between these patterns and the reasons that
might explain the relative immutability of socioeconomic problems in
some neighborhoods.
Abstract: This paper proposes a new approach to perform the
problem of real-time face detection. The proposed method combines
primitive Haar-Like feature and variance value to construct a new
feature, so-called Variance based Haar-Like feature. Face in image
can be represented with a small quantity of features using this
new feature. We used SVM instead of AdaBoost for training and
classification. We made a database containing 5,000 face samples
and 10,000 non-face samples extracted from real images for learning
purposed. The 5,000 face samples contain many images which have
many differences of light conditions. And experiments showed that
face detection system using Variance based Haar-Like feature and
SVM can be much more efficient than face detection system using
primitive Haar-Like feature and AdaBoost. We tested our method on
two Face databases and one Non-Face database. We have obtained
96.17% of correct detection rate on YaleB face database, which is
higher 4.21% than that of using primitive Haar-Like feature and
AdaBoost.
Abstract: Effective cooling of electronic equipment has emerged
as a challenging and constraining problem of the new century. In the
present work the feasibility and effectiveness of jet impingement
cooling on electronics were investigated numerically and
experimentally. Studies have been conducted to see the effect of the
geometrical parameters such as jet diameter (D), jet to target
spacing (Z) and ratio of jet spacing to jet diameter (Z/D) on the heat
transfer characteristics. The values of Reynolds numbers considered
are in the range 7000 to 42000. The results obtained from the
numerical studies are validated by conducting experiments. From the
studies it is found that the optimum value of Z/D ratio is 5. For a
given Reynolds number, the Nusselt number increases by about 28%
if the diameter of the nozzle is increased from 1mm to 2mm.
Correlations are proposed for Nusselt number in terms of Reynolds
number and these are valid for air as the cooling medium.
Abstract: SoftBoost is a recently presented boosting algorithm,
which trades off the size of achieved classification margin and
generalization performance. This paper presents a performance
evaluation of SoftBoost algorithm on the generic object recognition
problem. An appearance-based generic object recognition
model is used. The evaluation experiments are performed using
a difficult object recognition benchmark. An assessment with respect
to different degrees of label noise as well as a comparison to
the well known AdaBoost algorithm is performed. The obtained
results reveal that SoftBoost is encouraged to be used in cases
when the training data is known to have a high degree of noise.
Otherwise, using Adaboost can achieve better performance.
Abstract: This paper investigates the issue of building decision
trees from data with imprecise class values where imprecision is
encoded in the form of possibility distributions. The Information
Affinity similarity measure is introduced into the well-known gain
ratio criterion in order to assess the homogeneity of a set of
possibility distributions representing instances-s classes belonging to
a given training partition. For the experimental study, we proposed an
information affinity based performance criterion which we have used
in order to show the performance of the approach on well-known
benchmarks.
Abstract: Training neural networks to capture an intrinsic
property of a large volume of high dimensional data is a difficult
task, as the training process is computationally expensive. Input
attributes should be carefully selected to keep the dimensionality of
input vectors relatively small.
Technical indexes commonly used for stock market prediction
using neural networks are investigated to determine its effectiveness
as inputs. The feed forward neural network of Levenberg-Marquardt
algorithm is applied to perform one step ahead forecasting of
NASDAQ and Dow stock prices.
Abstract: The Multi-Layered Perceptron (MLP) Neural
networks have been very successful in a number of signal processing
applications. In this work we have studied the possibilities and the
met difficulties in the application of the MLP neural networks for the
prediction of daily solar radiation data. We have used the Polack-Ribière algorithm for training the neural networks. A comparison, in
term of the statistical indicators, with a linear model most used in
literature, is also performed, and the obtained results show that the
neural networks are more efficient and gave the best results.
Abstract: Primary barrier of membrane type LNG containment system consist of corrugated 304L stainless steel. This 304L stainless steel is austenitic stainless steel which shows different material behaviors owing to phase transformation during the plastic work. Even though corrugated primary barriers are subjected to significant amounts of pre-strain due to press working, quantitative mechanical behavior on the effect of pre-straining at cryogenic temperatures are not available. In this study, pre-strain level and pre-strain temperature dependent tensile tests are carried to investigate mechanical behaviors. Also, constitutive equations with material parameters are suggested for a verification study.
Abstract: The goal of this project is to design a system to
recognition voice commands. Most of voice recognition systems
contain two main modules as follow “feature extraction" and “feature
matching". In this project, MFCC algorithm is used to simulate
feature extraction module. Using this algorithm, the cepstral
coefficients are calculated on mel frequency scale. VQ (vector
quantization) method will be used for reduction of amount of data to
decrease computation time. In the feature matching stage Euclidean
distance is applied as similarity criterion. Because of high accuracy
of used algorithms, the accuracy of this voice command system is
high. Using these algorithms, by at least 5 times repetition for each
command, in a single training session, and then twice in each testing
session zero error rate in recognition of commands is achieved.
Abstract: Ren et al. presented an efficient carrier frequency offset
(CFO) estimation method for orthogonal frequency division multiplexing
(OFDM), which has an estimation range as large as the
bandwidth of the OFDM signal and achieves high accuracy without
any constraint on the structure of the training sequence. However,
its detection probability of the integer frequency offset (IFO) rapidly
varies according to the fractional frequency offset (FFO) change. In
this paper, we first analyze the Ren-s method and define two criteria
suitable for detection of IFO. Then, we propose a novel method for
the IFO estimation based on the maximum-likelihood (ML) principle
and the detection criteria defined in this paper. The simulation results
demonstrate that the proposed method outperforms the Ren-s method
in terms of the IFO detection probability irrespective of a value of
the FFO.
Abstract: This research was conducted for the first time at the
southeastern coasts of the Caspian Sea in order to evaluate the
performance of osteichthyes cooperatives through production (catch)
function. Using one of the indirect valuation methods in this research,
contributory factors in catch were identified and were inserted into
the function as independent variables. In order to carry out this
research, the performance of 25 Osteichthyes catching cooperatives
in the utilization year of 2009 which were involved in fishing in
Miankale wildlife refuge region. The contributory factors in catch
were divided into groups of economic, ecological and biological
factors. In the mentioned function, catch rate of the cooperative were
inserted into as the dependant variable and fourteen partial variables
in terms of nine general variables as independent variables. Finally,
after function estimation, seven variables were rendered significant at
99 percent reliably level. The results of the function estimation
indicated that human resource (fisherman quantity) had the greatest
positive effect on catch rate with an influence coefficient of 1.7 while
weather conditions had the greatest negative effect on the catch rate
of cooperatives with an influence coefficient of -2.07. Moreover,
factors like member's share, experience and fisherman training and
fishing effort played the main roles in the catch rate of cooperative
with influence coefficients of 0.81, 0.5 and 0.21, respectively.
Abstract: In the current research, neuro-fuzzy model and regression model was developed to predict Material Removal Rate in Electrical Discharge Machining process for AISI D2 tool steel with copper electrode. Extensive experiments were conducted with various levels of discharge current, pulse duration and duty cycle. The experimental data are split into two sets, one for training and the other for validation of the model. The training data were used to develop the above models and the test data, which was not used earlier to develop these models were used for validation the models. Subsequently, the models are compared. It was found that the predicted and experimental results were in good agreement and the coefficients of correlation were found to be 0.999 and 0.974 for neuro fuzzy and regression model respectively
Abstract: Using neural network we try to model the unknown function f for given input-output data pairs. The connection strength of each neuron is updated through learning. Repeated simulations of crisp neural network produce different values of weight factors that are directly affected by the change of different parameters. We propose the idea that for each neuron in the network, we can obtain quasi-fuzzy weight sets (QFWS) using repeated simulation of the crisp neural network. Such type of fuzzy weight functions may be applied where we have multivariate crisp input that needs to be adjusted after iterative learning, like claim amount distribution analysis. As real data is subjected to noise and uncertainty, therefore, QFWS may be helpful in the simplification of such complex problems. Secondly, these QFWS provide good initial solution for training of fuzzy neural networks with reduced computational complexity.
Abstract: The aim of this paper is to present the role of
myotonometry in assessment muscle viscoelasticity by measurement
of force index (IF) and stiffness (S) at thigh muscle groups. The
results are used for improve the muscle training. The method is based
on mechanic impulse on the muscle group, that involve a muscle
response like acceleration, speed and amplitude curves. From these
we have information about elasticity, stiffness beginning from
mechanic oscillations of muscle tissue. Using this method offer the
possibility for monitoring the muscle capacity for produce mechanic
energy, that allows a efficiency of movement with a minimal tissue
deformation.
Abstract: In this paper back-propagation artificial neural
network (BPANN) with Levenberg–Marquardt algorithm is
employed to predict the limiting drawing ratio (LDR) of the deep
drawing process. To prepare a training set for BPANN, some finite
element simulations were carried out. die and punch radius, die arc
radius, friction coefficient, thickness, yield strength of sheet and
strain hardening exponent were used as the input data and the LDR
as the specified output used in the training of neural network. As a
result of the specified parameters, the program will be able to
estimate the LDR for any new given condition. Comparing FEM and
BPANN results, an acceptable correlation was found.
Abstract: In this paper we present an adaptive method for image
compression that is based on complexity level of the image. The
basic compressor/de-compressor structure of this method is a multilayer
perceptron artificial neural network. In adaptive approach
different Back-Propagation artificial neural networks are used as
compressor and de-compressor and this is done by dividing the
image into blocks, computing the complexity of each block and then
selecting one network for each block according to its complexity
value. Three complexity measure methods, called Entropy, Activity
and Pattern-based are used to determine the level of complexity in
image blocks and their ability in complexity estimation are evaluated
and compared. In training and evaluation, each image block is
assigned to a network based on its complexity value. Best-SNR is
another alternative in selecting compressor network for image blocks
in evolution phase which chooses one of the trained networks such
that results best SNR in compressing the input image block. In our
evaluations, best results are obtained when overlapping the blocks is
allowed and choosing the networks in compressor is based on the
Best-SNR. In this case, the results demonstrate superiority of this
method comparing with previous similar works and JPEG standard
coding.
Abstract: A new conceptual architecture for low-level neural
pattern recognition is presented. The key ideas are that the brain
implements support vector machines and that support vectors are
represented as memory patterns in competitive queuing memories. A
binary classifier is built from two competitive queuing memories
holding positive and negative valence training examples respectively.
The support vector machine classification function is calculated in
synchronized evaluation cycles. The kernel is computed by bisymmetric
feed-forward networks feed by sensory input and by
competitive queuing memories traversing the complete sequence of
support vectors. Temporary summation generates the output
classification. It is speculated that perception apparatus in the brain
reuses structures that have evolved for enabling fluent execution of
prepared action sequences so that pattern recognition is built on
internalized motor programmes.