Kalman filter decision systems for debris flow hazard assessment

Abstract

This paper aims to develop a model for debris flow hazard assessment, since Taiwan is a mountainous country subject to bouts of heavy rainfall during the rainy and typhoon seasons and is thus frequently subject to landslide disasters. The database used is comprised of information from actual cases that occurred in areas of Hualien in eastern Taiwan during 2007 and 2008. The Kalman filter model is utilized to assess the occurrence of debris flows from computed indexes, to verify modeling errors. Comparisons are made between two models to determine which one is better in practical applications. The efficiency of the Kalman filter decision system has been proved, showing smaller average relative error and correspondingly larger ratio of success on debris flow assessment when compared to the neural network model. The relative error is calculated between the differences in the ratio of model outputs and that of actual occurrences. Such an error index can be decreased from 4.65 to 3.39% by introducing the concept of geographic divisions to the Kalman filter model, which demonstrates its ability to forecast the occurrence of debris flows for the coming year.

Piecewise statistical programming for workforce allocation: the significance of Taiwanese cadres in China

“…The quadratic curve method is a kind of prediction method (Lin and Chen, 2009;Lin, 2011;Lin et al, 2012) of extrapolation curve that is being used to survey the trend changes of time-series observation data over time which would likely show from high level to low level, and then, back to high level (or from low level to high level, and then back to low level). As the scatter plots of time-series observation data shows the changes in parabolic shape, it has been named as the "quadratic (parabolic) prediction model" (Akopyan and Zaslavsky, 2007).…”

A Framework for Hazard Assessment of Debrisflow Accumulation Areas

“…The most common empirical relationships estimate the maximum run-out distance of debris flow. In recent years, data mining has been utilized for assessing debris flow hazards (Lin et al, 2012;Liang et al, 2012). With the development of rheological and physical-mathematical based modeling of debris flows, numerical techniques and geographical information science, quantitative approaches are widelyused to assess the hazard of debris-flow.…”

Detecting GSM churners by using Euclidean Indexing HDMR

“…Another study proposes a probabilistic model based on historical events to be used to assess flood risk in Taiwan [14]. The Kalman filter model is used to determine the occurrence of debris flows [15]. Two studies about forecasting fresh food sales are based on the usage of logistic regression, moving average and BPNN methods [16,17].…”