期刊名称：International Journal of Advanced Computer Science and Applications(IJACSA)
印刷版ISSN：2158-107X
电子版ISSN：2156-5570
出版年度：2019
卷号：10
期号：9
页码：352-359
出版社：Science and Information Society (SAI)
摘要：Several regions of the world have been affected by
one of the natural disasters named as flash floods. Many villagers
who live near stream or dam, they suffer a lot in terms of
property, cattle and human lives loss. Conventional early
warning systems are not up to the mark for the early warning
announcements. Diversified approaches have been carried out
for the identification of flash floods with less false alarm rate.
Forecasting approaches includes some errors and ambiguity due
to the incompetent processing algorithms and measurement
readings. Process variables like stream flow, water level, water
color, precipitation velocity, wind speed, wave’s pattern and
cloud to ground (CG) flashes have been measured for the robust
identification of flash floods. A vibrant competent algorithm
would be required for the investigation of flash floods with less
false alarm rate. In this research paper classifiers have been
applied on the collected data set so that any researcher could
easily know that which classifier is competent and can be further
enhanced by combining it with other algorithms. A novel
comprehensive parametric comparison has been performed to
investigate the classification accuracy for the robust classification
of false alarms. For the better accuracy more than one process
variables have been measured but still contained some false
alarm. Appropriate combination of sensor was integrated to
increase the accuracy in results as multi-modal sensing device
has been designed to collect the data. Linear discriminant
analysis, logistic regression, quadratic support vector machine, knearest
neighbor and Ensemble bagged tree have been applied to
the collected data set for the data classification. Results have
been obtained in the MATLAB and discussed in detail in the
research paper. The worst accuracy of the classification (62%)
has been achieved by the coarse k-NN classifier that means
coarse k-NN produced 38% false negative rate that is not
acceptable in the case of forecasting. Ensemble bagged trees
produced best classification results as it achieved 99 % accuracy
and 1% error rate. Furthermore, according to the
comprehensive parametric comparison of regression models
Quadratic SVM found to be the worst with mean square error of
0.5551 and time elapsed 13.159 seconds. On the other hand,
Exponential Gaussian process regression performed better than
the other existing approaches with the minimum root mean
squared error of 0.0002 and prediction speed of 35000
observations per second.
