發(fā)布時(shí)間：2018-06-23 09:47

  本文選題：機(jī)器學(xué)習(xí) + 隨機(jī)森林�。� 參考：《上海應(yīng)用技術(shù)大學(xué)》2017年碩士論文

【摘要】：近年來(lái),城市空氣污染問(wèn)題愈發(fā)突出,大量的汽車(chē)尾氣、工業(yè)廢氣、揚(yáng)塵等直接排放到城市環(huán)境中,遠(yuǎn)遠(yuǎn)超過(guò)了城市環(huán)境的自?xún)裟芰?導(dǎo)致空氣質(zhì)量下降,直接危及到城市居民的健康和安全,城市空氣質(zhì)量問(wèn)題已經(jīng)引起了中國(guó)社會(huì)各界的高度重視。為了有效地控制空氣污染,提高城市的空氣質(zhì)量,首先一定要對(duì)城市環(huán)境空氣質(zhì)量進(jìn)行科學(xué)合理的評(píng)價(jià),使城市居民以及環(huán)境保護(hù)部門(mén)更加客觀地了解城市環(huán)境的空氣質(zhì)量,做出合理的生活安排及科學(xué)的防治措施。所以城市環(huán)境空氣質(zhì)量評(píng)價(jià)對(duì)于防治空氣污染發(fā)揮著重要的作用。隨著大數(shù)據(jù)時(shí)代的來(lái)臨以及人工智能的興起,傳統(tǒng)的評(píng)價(jià)方法已經(jīng)不能滿(mǎn)足對(duì)海量傳感器數(shù)據(jù)智能化處理的需求,大量學(xué)者開(kāi)始基于大數(shù)據(jù),利用智能方法來(lái)評(píng)價(jià)城市環(huán)境的空氣質(zhì)量。機(jī)器學(xué)習(xí)是人工智能的分支,本文將機(jī)器學(xué)習(xí)引入到城市環(huán)境空氣質(zhì)量評(píng)價(jià)中,利用機(jī)器學(xué)習(xí)中的隨機(jī)森林算法來(lái)評(píng)價(jià)城市環(huán)境的空氣質(zhì)量,通過(guò)對(duì)隨機(jī)森林模型訓(xùn)練,找到多種空氣污染物與空氣質(zhì)量等級(jí)之間的內(nèi)在映射關(guān)系,建立隨機(jī)森林評(píng)價(jià)模型,提高評(píng)價(jià)科學(xué)性和魯棒性。本文在仿真實(shí)驗(yàn)時(shí),將隨機(jī)森林評(píng)價(jià)模型與支持向量機(jī),樸素貝葉斯和K最鄰近模型進(jìn)行對(duì)比,仿真采用的數(shù)據(jù)為上海市2013年到2015年的部分空氣質(zhì)量真實(shí)數(shù)據(jù),仿真實(shí)驗(yàn)得到了較好的效果,實(shí)驗(yàn)結(jié)果表明評(píng)價(jià)方法效果最好,準(zhǔn)確性最高可達(dá)99.69%,同時(shí)本文對(duì)隨機(jī)森林模型的性能進(jìn)行了深入分析,進(jìn)一步驗(yàn)證了該評(píng)價(jià)方法的適用性與穩(wěn)定性,從分析可以看出本文模型的泛化誤差對(duì)特征變量個(gè)數(shù)不是很敏感,并且在準(zhǔn)確性與時(shí)間復(fù)雜度之間有較好折衷,可以用于準(zhǔn)確有效的評(píng)價(jià)城市環(huán)境的空氣質(zhì)量。
[Abstract]:In recent years, the problem of urban air pollution has become more and more prominent. A large number of automobile exhaust, industrial waste gas and dust are discharged directly into the urban environment, which far exceeds the self purification capacity of the urban environment, which leads to the decline of air quality, which directly endangers the health and safety of urban residents. The problem of urban air quality has already caused the high social circles in China. In order to effectively control air pollution and improve the air quality of the city, first of all, we must make a scientific and rational evaluation of the air quality of the urban environment, so that the urban residents and the environmental protection departments can understand the air quality of the urban environment more objectively and make reasonable living arrangements and scientific prevention measures. Air quality assessment plays an important role in preventing air pollution. With the advent of the era of large data and the rise of artificial intelligence, the traditional evaluation method can not meet the demand for intelligent processing of mass sensor data. A large number of scholars begin to use intelligent methods to evaluate the air of the urban environment based on large data. Quality. Machine learning is the branch of artificial intelligence. This paper introduces machine learning into urban environmental air quality evaluation, uses random forest algorithm in machine learning to evaluate the air quality of urban environment, and through the training of random forest model, the internal mapping relationship between air pollution and air quality is found. The stochastic forest evaluation model is established to improve the scientific and robust evaluation. In the simulation experiment, the stochastic forest evaluation model is compared with the support vector machine, the naive Bias and the K nearest model. The simulation data are the real data of the air quality in Shanghai from 2013 to 2015, and the simulation experiment is better. The results show that the evaluation method has the best effect and the maximum accuracy can reach 99.69%. At the same time, the performance of the random forest model is analyzed in this paper, and the applicability and stability of the evaluation method are further verified. There is a good trade-off between accuracy and time complexity, which can be used to evaluate the air quality of urban environment accurately and effectively.
【學(xué)位授予單位】：上海應(yīng)用技術(shù)大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類(lèi)號(hào)】：X823

【參考文獻(xiàn)】

相關(guān)期刊論文 前4條

1 薛志鋼;郝吉明;陳復(fù);柴發(fā)合;;歐美發(fā)達(dá)國(guó)家大氣污染控制經(jīng)驗(yàn)[J];杭州(周刊);2016年05期

2 姜新華;劉霞;薛河儒;張存厚;;基于逐步回歸的空氣質(zhì)量影響因素分析——以呼和浩特市區(qū)為例[J];內(nèi)蒙古農(nóng)業(yè)大學(xué)學(xué)報(bào)(自然科學(xué)版);2015年02期

3 李鵬;;能源消費(fèi)與我國(guó)的二氧化硫排放——兼論人口規(guī)模及技術(shù)進(jìn)步對(duì)二氧化硫排放的影響[J];西北人口;2014年04期

4 廖銀念;蘇玉紅;艾尼瓦爾·買(mǎi)買(mǎi)提;;城市空氣質(zhì)量的模糊綜合評(píng)價(jià)——以烏魯木齊市為例[J];北方環(huán)境;2011年11期

相關(guān)博士學(xué)位論文 前1條

1 劉杰;北京大氣污染物時(shí)空變化規(guī)律及評(píng)價(jià)預(yù)測(cè)模型研究[D];北京科技大學(xué);2015年

相關(guān)碩士學(xué)位論文 前6條

1 趙敏;基于層次分析法的地下水超采區(qū)劃分及壓采效果評(píng)價(jià)[D];太原理工大學(xué);2016年

2 周家?guī)?基于多元統(tǒng)計(jì)和智能算法的上海市空氣質(zhì)量指數(shù)評(píng)價(jià)分析[D];蘭州大學(xué);2016年

3 丁里;基于機(jī)器學(xué)習(xí)的P2P網(wǎng)絡(luò)流分類(lèi)研究[D];江南大學(xué);2015年

4 王露云;中國(guó)31個(gè)主要城市空氣質(zhì)量評(píng)價(jià)及主要污染物濃度預(yù)測(cè)[D];重慶師范大學(xué);2014年

5 張良;石家莊市“十一五”期間環(huán)境空氣質(zhì)量變化趨勢(shì)分析及預(yù)測(cè)研究[D];河北科技大學(xué);2013年

6 陳瑋;基于灰色聚類(lèi)與模糊綜合評(píng)判的空氣質(zhì)量評(píng)價(jià)[D];華東師范大學(xué);2012年

，

本文編號(hào)：2056713