基於深度學習的道路短期交通狀態時空序列預測
崔建勛 等
- 出版商: 電子工業
- 出版日期: 2022-04-01
- 售價: $588
- 貴賓價: 9.5 折 $559
- 語言: 簡體中文
- 頁數: 296
- ISBN: 7121430193
- ISBN-13: 9787121430190
-
相關分類:
DeepLearning、Machine Learning
立即出貨 (庫存 < 3)
買這商品的人也買了...
-
$2,010$1,910 -
$480$379 -
$352密碼學 (C\C++語言實現原書第2版)
-
$594$564 -
$800$760 -
$403S7-1200/1500 PLC 應用技術
-
$301特徵工程入門與實踐 (Feature Engineering Made Easy)
-
$332了不起的 Markdown
-
$880$748 -
$454聯邦學習
-
$774$735 -
$2,250$2,138 -
$505深入淺出 Embedding:原理解析與應用實踐
-
$620$527 -
$699$594 -
$2,680$2,546 -
$719$683 -
$620$484 -
$500$395 -
$594$564 -
$380$323 -
$305知識圖譜:方法、工具與案例
-
$454從零開始大模型開發與微調:基於 PyTorch 與 ChatGLM
-
$454精通推薦算法:核心模塊 + 經典模型 + 代碼詳解
-
$422ChatGLM3大模型本地化部署、應用開發與微調
相關主題
商品描述
這本書系統闡述了深度學習方法論在道路短期交通狀態時空序列預測領域的最新研究成果。需要著重說明以下幾點:(1)領域限定在了道路交通,因為交通是個大系統,存在著航空、水運、道路等多種運輸方式,而本書所闡述的研究均是針對道路交通領域的數據以及面向道路交通領域的應用;(2)本書所討論的研究問題是道路短期交通狀態時空序列預測問題,該問題是時空數據挖掘領域中時空預測問題的一個重要子集,在本書的第1章中將會對這個問題進行數學上的形式化定義;(3)本書針對道路短期交通狀態時空序列預測問題的討論,完全是基於深度學習的方法論,所參考的文獻絕大部分發表於2017年以後,並不涵蓋前人對該研究問題所採用的全部方法論(如ARIMA,卡爾曼濾波、SVR等)。
目錄大綱
目錄
□ □ 章道路短期交通狀態時空序列預測□□....................................................001
1.1 時空數據...............................................................................................................001
1.□ 時空數據挖掘.......................................................................................................00□
1.3 道路短期交通狀態時空序列預測.......................................................................003
1.3.1 問題描述..................................................................................................003
1.3.□ 核心挑戰..................................................................................................005
1.3.3 問題分類..................................................................................................007
1.4 道路短期交通狀態時空序列預測研究概要性綜述...........................................01□
1.5 基於深度學□□道路短期交通狀態時空序列預測建模一般性框架................014
1.6 本章小結...............................................................................................................015
□ □ 篇基於深度學□□網格化道路交通狀態時空序列預測
第□ 章基於□D 圖像卷積神經網絡的時空相關性建模...................................018
□.1 ST-ResNet .............................................................................................................0□0
□.1.1 問題提出..................................................................................................0□0
□.1.□ 歷史交通狀態切片數據的獲取...............................................................0□0
□.1.3 預測模型..................................................................................................0□□
□.1.4 訓練算法..................................................................................................0□6
□.□ MDL......................................................................................................................0□7
□.□.1 問題提出..................................................................................................0□7
□.□.□ 預測模型..................................................................................................0□9
□.□.3 訓練算法..................................................................................................035
□.3 MF-STN ................................................................................................................036
□.3.1 問題提出..................................................................................................037
□.3.□ 預測模型..................................................................................................037
□.3.3 訓練算法..................................................................................................040
□.4 DeepLGR[□3] ..........................................................................................................04□
□.4.1 問題提出..................................................................................................043
□.4.□ 預測模型..................................................................................................043
□.4.3 模型小結..................................................................................................048
□.5 ST-NASNet ...........................................................................................................048
□.5.1 問題提出..................................................................................................051
□.5.□ 預測模型..................................................................................................051
□.5.3 訓練算法..................................................................................................054
□.6 本章小結...............................................................................................................055
第3 章基於□D 圖像卷積與循環神經網絡相結合的時空相關性建模.......057
3.1 STDN[□5]................................................................................................................058
3.1.1 問題提出..................................................................................................059
3.1.□ 預測模型..................................................................................................059
3.1.3 訓練算法..................................................................................................066
3.□ ACFM[□6] ...............................................................................................................067
3.□.1 問題提出..................................................................................................067
3.□.□ 預測模型..................................................................................................068
3.□.3 模型拓展..................................................................................................073
3.□.4 訓練算法..................................................................................................075
3.3 PredRNN[□7] ..........................................................................................................076
3.4 PredRNN++[□8] ......................................................................................................081
3.4.1 模型架構..................................................................................................08□
3.4.□ Casual-LSTM............................................................................................083
3.4.3 GHU..........................................................................................................084
3.5 MIM[□9]..................................................................................................................084
3.6 SA-ConvLSTM[30].................................................................................................088
3.6.1 模型背景..................................................................................................089
3.6.□ 模型構造..................................................................................................090
3.7 本章小結...............................................................................................................09□
第4 章基於3D 圖像卷積的時空相關性建模.....................................................094
4.1 問題提出...............................................................................................................095
4.□ 預測模型...............................................................................................................095
4.□.1 近期時空相關性捕獲模塊.......................................................................096
4.□.□ 短期時空相關性捕獲模塊.......................................................................098
4.□.3 特徵融合模塊...........................................................................................099
4.□.4 預測模塊..................................................................................................099
4.□.5 損失函數..................................................................................................099
4.3 訓練算法...............................................................................................................100
4.4 本章小結...............................................................................................................100
第□ 篇基於深度學□□拓撲化道路交通狀態時空序列預測
第5 章基於1D 圖像卷積與卷積圖神經網絡相結合的時空相關性建模..10□
5.1 STGCN[3□] .............................................................................................................10□
5.1.1 問題提出..................................................................................................10□
5.1.□ 模型建立..................................................................................................103
5.□ TSSRGCN[33] ........................................................................................................105
5.□.1 問題提出..................................................................................................106
5.□.□ 模型建立..................................................................................................106
5.3 Graph Wave Net[34]................................................................................................11□
5.3.1 問題提出..................................................................................................11□
5.3.□ 模型建立..................................................................................................113
5.4 ASTGCN[35] ..........................................................................................................116
5.4.1 問題提出..................................................................................................116
5.4.□ 模型建立..................................................................................................117
5.5 本章小結...............................................................................................................1□3
第6 章基於循環與卷積圖神經網絡相結合的時空相關性建模....................1□4
6.1 AGC-Seq□Seq[36]...................................................................................................1□4
6.1.1 問題提出..................................................................................................1□5
6.1.□ 模型建立..................................................................................................1□5
6.□ DCGRU[37] ............................................................................................................1□9
6.□.1 問題提出..................................................................................................130
6.□.□ 模型建立..................................................................................................130
6.3 T-MGCN[38] ...........................................................................................................13□
6.3.1 問題提出..................................................................................................13□
6.3.□ 模型建立..................................................................................................133
6.4 GGRU[39] ...............................................................................................................138
6.4.1 符號定義..................................................................................................139
6.4.□ GaAN 聚合器...........................................................................................140
6.4.3 GGRU 循環單元......................................................................................141
6.4.4 基於Encoder-Decoder 架構和GGRU 的交通狀態時空預測網絡........141
6.5 ST-MetaNet[40].......................................................................................................14□
6.5.1 問題提出..................................................................................................143
6.5.□ 模型建立..................................................................................................143
6.6 本章小結...............................................................................................................147
第7 章基於Self-Attention 與卷積圖神經網絡相結合的時空相關性建模....149
7.1 GMAN[41] ..............................................................................................................150
7.1.1 問題提出..................................................................................................150
7.1.□ 模型建立..................................................................................................150
7.□ ST-GRAT[4□] ..........................................................................................................157
7.□.1 問題提出..................................................................................................157
7.□.□ 模型建立..................................................................................................158
7.3 STTN[43] ................................................................................................................163
7.3.1 問題提出..................................................................................................163
7.3.□ 模型建立..................................................................................................164
7.4 STGNN[44] .............................................................................................................169
7.4.1 問題提出..................................................................................................169
7.4.□ 模型建立..................................................................................................169
7.5 本章小結...............................................................................................................173
第8 章基於卷積圖神經網絡的時空相關性同步建模......................................174
8.1 MVGCN[45] ...........................................................................................................175
8.1.1 問題提出..................................................................................................176
8.1.□ 模型建立..................................................................................................177
8.□ STSGCN[46] ...........................................................................................................180
8.□.1 問題提出..................................................................................................180
8.□.□ 模型建立..................................................................................................180
8.3 本章小結...............................................................................................................186
第3 篇深度學習相關基本理論
第9 章全連接神經網絡.............................................................................................190
9.1 理論介紹...............................................................................................................190
9.□ 本章小結...............................................................................................................19□
□ □0 章卷積神經網絡...............................................................................................193
10.1 二維卷積神經網絡(□D CNN).......................................................................193
10.□ 一維卷積和三維卷積神經網絡(1D 和3D CNN) ........................................198
10.3 擠壓和激勵卷積網絡(Squeeze and Excitation Networks)............................199
10.4 殘差連接網絡(ResNet) .................................................................................□01
10.5 因果卷積(Casual CNN).................................................................................□0□
10.6 膨脹卷積(Dilated Convolution) ....................................................................□03
10.7 可□形卷積(Deformable Convolution) .........................................................□04
10.8 可分離卷積(Separable Convolution) ............................................................□06
10.9 亞像素卷積(SubPixel Convolution)..............................................................□07
10.10 本章小結...........................................................................................................□08
□ □1 章循環神經網絡................................................................................................□10
11.1 標準循環神經網絡(RNN).............................................................................□11
11.□ 雙向循環神經網絡(Bi-RNN)........................................................................□11
11.3 深度循環神經網絡(Deep RNN) ...................................................................□1□
11.4 長短期記憶神經網絡(LSTM)[60] ..................................................................□13
11.5 門控循環單元(GRU).....................................................................................□15
11.6 ConvLSTM .........................................................................................................□16
11.7 本章小結.............................................................................................................□17
□ □□ 章卷積圖神經網絡...........................................................................................□18
1□.1 譜域圖卷積[66] ....................................................................................................□□0
1□.1.1 拓撲圖數據上的捲積操作推導.............................................................□□0
1□.1.□ 切比雪夫多項式捲積.............................................................................□□5
1□.1.3 圖卷積網絡(Graph Convolutional Networks,GCN).......................□□6
1□.1.4 擴散卷積(Diffusion Convolution).....................................................□□6
1□.□ 空間域圖卷積.....................................................................................................□□8
1□.□.1 頂點域圖卷積特徵聚合器的一般性定義.............................................□□8
1□.□.□ GraphSAGE[71]........................................................................................□□9
1□.□.3 GAT.........................................................................................................□3□
1□.3 本章小結.............................................................................................................□35
□ □3 章註意力機制(Attention).........................................................................□36
13.1 Encoder-Decoder 模型[75-77] ................................................................................□36
13.□ 基於注意力機制的Encoder-Decoder 模型[78-80] ...............................................□38
13.3 廣義注意力機制[81-83] .........................................................................................□40
13.4 多頭注意力機制(Multi-Head Attention)[84-87] ...............................................□41
13.5 自註意力機制(Self-Attention)[88-91] ..............................................................□4□
13.6 Encoder-Decoder 架構的□體及訓練方法........................................................□45
13.7 本章小結.............................................................................................................□49
□ □4 章Transformer[74,94-97] ....................................................................................□50
14.1 模型介紹.............................................................................................................□51
14.□ 本章小結.............................................................................................................□54
□ □5 章深度神經網絡訓練技巧.............................................................................□55
15.1 Batch Normalization(BN) ..............................................................................□55
15.□ Layer Normalization(LN)[99] ..........................................................................□6□
15.3 本章小結.............................................................................................................□63
□ □6 章矩陣分解(Matrix Factorization)[100] ................................................□64
16.1 理論介紹.............................................................................................................□64
16.□ 本章小結.............................................................................................................□67
後記.......................................................................................................................................□68
參考文獻..............................................................................................................................□70