發(fā)布時間：2021-05-21 23:40

　　隨著中國經(jīng)濟(jì)的快速發(fā)展,貨運市場持續(xù)增長,公路、鐵路、水路等貨運量不斷增加。然而,貨運量信息是很難監(jiān)控的,因為它數(shù)據(jù)來源較廣,需要將眾多分散信息匯總、過濾并儲存在數(shù)據(jù)庫中。因此,本文利用網(wǎng)絡(luò)爬蟲技術(shù)對公路、水路貨運市場的運量數(shù)據(jù)進(jìn)行了獲取和匯總,運用二次曲線等模型對缺失數(shù)據(jù)進(jìn)行了修復(fù),并基于歷史數(shù)據(jù)采用指數(shù)平滑模型對未來運量走勢進(jìn)行了預(yù)測。本文首先利用八爪魚網(wǎng)絡(luò)爬蟲工具從不同網(wǎng)站上獲取貨運市場的運量信息,按周、月、季、年、運輸方式、地區(qū)和貨運量進(jìn)行細(xì)分,以地區(qū)為單位,對同種運輸方式下不同時間段內(nèi)的運量數(shù)據(jù)進(jìn)行匯總,同時探索出數(shù)據(jù)中所有的缺失值之后以華北、山東、華中地區(qū)的貨運市場的運量信息為例,進(jìn)行數(shù)據(jù)修復(fù)。選取缺失數(shù)據(jù)所在年份的其他數(shù)據(jù),利用二次曲線模型、指數(shù)模型、對數(shù)模型等進(jìn)行曲線擬合,根據(jù)不同模型的R方值,選取擬合效果最好的模型,確定模型參數(shù)并計算缺失值,依次修復(fù)全部缺失值。最后,以2016年、2017年、2018年的數(shù)據(jù)為基礎(chǔ),預(yù)測2019年前22周的數(shù)據(jù)。采用指數(shù)平滑法建立了上述三個地區(qū)的未來貨運量預(yù)測模型,考慮到數(shù)據(jù)明顯的季節(jié)性,采用三次指數(shù)平滑法（霍爾特冬季指數(shù)平滑法）對比... 

【文章來源】：北京交通大學(xué)北京市 211工程院校 教育部直屬院校

【文章頁數(shù)】：81 頁

【學(xué)位級別】：碩士

【文章目錄】：
Acknowledgement
中文摘要
ABSTRACT
1 INTRODUCTION
    1.1 MOTIVATION
    1.2 OBJECTIVES
    1.3 SCOPE OF RESEARCH
    1.4 RESEARCH SIGNIFICANCE
    1.5 THESIS OUTLINE
2 LITERATURE REVIEW
    2.1 TIME SERIES STUDIES
    2.2 SIMPLE EXPONENTIAL SMOOTHING MODEL
        2.2.1. Mathematical Formulation
    2.3. Measuring Forecast Error
        2.3.1 Choosing the Best Value for Smoothing Constant
    2.4 WEB CRAWLER
    2.5 REVIEW OF THE DATA SCRAPING TOOLS THAT NEITHER REQUIRE PROGRAMMING NOR CODING
        2.5.1 Octoparse
        2.5.2 Outwit hub
        2.5.3. Visual scraper
        2.5.4 Helium scraper
    2.6 WHAT IS SQL
3 RESEARCH METHODOLOGY
    3.1 OCTOPARSE OVERVIEW
        3.1.1 Installation
        3.1.2 Features
        3.1.3 Setting up basic information
        3.1.4 Morkflow Design
        3.1.5 Extraction Options
    3.2 IMPORT OF DATA EXTRACTED TO DATABASE
        3.2.1 Data and Database
        3.2.2 Creating table
    3.3 ACCESSING DATA
    3.4 MODEL FORMULATION
        3.4.1 Problem Definition
        3.4.2 Data repair
        3.4.3 The mean of adjacent points
    3.5 QUADRATIC CURVE MODEL
    3.6 APPLICATION OF SPSS TO CALCULATE
    3.7 EXPONENTIAL SMOOTHING MODEL
        3.7.1 Parameters for the model:
4 DATA INSPECTION AND ANALYSIS
    4.1 DATA SOURCES
    4.2 TRAFFIC FREIGHT VOLUME
    4.3 SELECTION OF MODEL FOR NORTH CHINA,SHANDONG PENINSULA AND HUAZHONG
    4.4 QUADRATIC CURVE MODEL APPLICATION
    4.5 ANALYSIS OF THE QUADRATIC CURVE MODEL
    4.6 REPLACE MISSING DATA BEFORE FORECAST
    4.7 DEFINE DATE FROM DATA
    4.8 PREDICTED CRITERIA FOR CUBIC EXPONENTIAL SMOOTHING
    4.9 THE AVERAGE ABSOLUTE ERROR(MAE)AND ROOT MEAN SQUARE ERROR(RMSE)
    4.10 COMPARISON OF THE PERFORMANCE MEASURES
    4.11 SUMMARY
5 CONCLUSIONS
REFERENCES
APPENDIX
AUTHOR PROFILE AND RESEARCH ACHIEVEMENTS OBTAINED DURING THE STUDYFOR A MASTER'S / DOCTORAL DEGREEIPUBLICATIONS
DATASET FOR THE MASTER'S THESIS



本文編號：3200573