Muutke küpsiste eelistusi

Flood Forecasting Using Artificial Neural Networks [Pehme köide]

(UNESCO-IHE Institute for Water Education, Delft, the Netherlands)
  • Formaat: Paperback / softback, 112 pages, kõrgus x laius: 234x156 mm, kaal: 210 g
  • Ilmumisaeg: 01-Jan-2003
  • Kirjastus: A A Balkema Publishers
  • ISBN-10: 9058096319
  • ISBN-13: 9789058096319
Teised raamatud teemal:
  • Pehme köide
  • Hind: 128,81 €*
  • * hind on lõplik, st. muud allahindlused enam ei rakendu
  • Tavahind: 171,75 €
  • Säästad 25%
  • Raamatu kohalejõudmiseks kirjastusest kulub orienteeruvalt 3-4 nädalat
  • Kogus:
  • Lisa ostukorvi
  • Tasuta tarne
  • Tellimisaeg 2-4 nädalat
  • Lisa soovinimekirja
Flood Forecasting Using Artificial Neural NetworksSuurem pilt
  • Formaat: Paperback / softback, 112 pages, kõrgus x laius: 234x156 mm, kaal: 210 g
  • Ilmumisaeg: 01-Jan-2003
  • Kirjastus: A A Balkema Publishers
  • ISBN-10: 9058096319
  • ISBN-13: 9789058096319
Teised raamatud teemal:
Püsilink: https://www.kriso.ee/db/9789058096319.html
Märksõnad:
Varoonchotikul (a civil engineer in Thailand) proposes a solution to the extrapolation problem, one of the most serious limitations to the Artificial Neural Network approach to modeling rainfall-runoff and predicting flooding. Focusing on the range of activation functions for each node, the proposed method seeks to create space for an Artificial Neural network to extrapolate. Case studies document the success of this approach in the Silk Stream, the Thrushel and Brue Rivers, and elsewhere. There is no index. Annotation (c) Book News, Inc., Portland, OR (booknews.com)

Flood disasters continue to occur in many countries in the world and cause tremendous casualties and property damage. To mitigate the effects of floods, a range of structural and non-structural measures have been employed including dykes, channelling, flood-proofing property, land-use regulation and flood warning schemes. Such schemes can include the use of Artificial Neural Networks (ANN) for modelling the rainfall run-off process as it is a quick and flexible approach which gives very promising results. However, the inability of ANN to extrapolate beyond the limits of the training range is a serious limitation of the method, and this book examines ways of side-stepping or solving this complex issue.
1 Introduction 1(10)
1.1 Flood forecasting
1(7)
1.2 Aims of this study
8(2)
1.2.1. The extrapolation problem
8(1)
1.2.2. Network architecture
8(1)
1.2.3. Choice of inputs
9(1)
1.3 Thesis outline
10(1)
2 Artificial Neural Networks 11(12)
2.1 Biological Neuron
11(1)
2.2 What is an Artificial Neural Network
12(1)
2.3 Multilayer Perceptron
13(2)
2.4 Delta learning rule for Feed-Forward Multilayer Perceptron
15(2)
2.5 Activation functions
17(4)
2.5.1 The unipolar binary function or sigmoid function (S)
17(1)
2.5.2 The bipolar binary function (B)
18(1)
2.5.3 The hyperbolic tangent function (T)
19(1)
2.5.4 The linear function (L)
20(1)
2.6 Recurrent Multilayer Perceptron
2.6.1 Elman Recurrent Network (ERN)
21(1)
2.6.2 Jordan Recurrent Network (JRN)
22(1)
3 Preliminary considerations 23(6)
3.1 Introduction
23(1)
3.2 The choice of a suitable time interval
23(3)
3.3 The choice of goodness-of-fit indices
26(3)
4 Extrapolation management for Artificial Neural Network models of Rainfall-Runoff relationships 29(18)
4.1 Introduction
29(1)
4.2 Standardisation: key to managing the extrapolation problem
30(1)
4.3 Standardised range of other activation functions
31(1)
4.4 Case studies without including the EMF in the raw data
32(4)
4.4.1 Silk Stream catchment
32(3)
4.4.2 Dollis Brook catchment
35(1)
4.4.3 Thrushel river
35(1)
4.4.4 Dart river
35(1)
4.5 Case studies with including EMF in the raw data
36(3)
4.6 Summary
39(1)
4.7 Case studies with different activation functions
39(4)
4.8 Different values of the factor K2
43(3)
4.9 Summary
46(1)
5 Recurrent Neural Networks 47(14)
5.1 Introduction
47(1)
5.2 Feed forward network
48(5)
5.2.1 Case I: the Brue River
48(1)
5.2.2 Case II: the Thrushel River
49(2)
5.2.3 Case III: the Chumporn River
51(2)
5.3 Elman recurrent network
53(3)
5.4 Jordan recurrent network
56(3)
5.4.1 Case I: the Brue River
56(1)
5.4.2 Case II: the Thrushel River
57(2)
5.5 Jordan recurrent network with Future Errors
59(1)
5.6 Concluding Remarks
60(1)
6 Choice of input 61(20)
6.1 Introduction
61(1)
6.2 Average rainfall or Distributed rainfall
61(4)
6.3 Forward shift operator
65(1)
6.4 Implementation of the FDTF Method
66(13)
6.4.1 Case I: Brue river
68(6)
6.4.2 Case II: Mole river
74(5)
6.5 Discussions and Conclusions
79(2)
7 Conclusions and recommendations 81(4)
8 Samenvatting 85(4)
9 References 89(12)
I Data used for the study
95(6)
1.1. The Brue river
95(1)
1.2. The Chumporn river
96(1)
1.3. The Dart river
97(1)
1.4. The Dollis Brook river
97(1)
1.5. The River Mole
98(1)
1.6. The Silk Stream river
99(1)
1.7. The Thrushel river
99(1)
1.8. Estimated Maximum Flood (EMF) of the Dart and Thrushel rivers
99(2)
Curriculum vitae 101
P Varoonchotikul