本文選題：數(shù)據(jù)挖掘 + 決策樹��； 參考：《江西理工大學(xué)》2015年碩士論文

【摘要】：中國是地質(zhì)災(zāi)害的多發(fā)國家,尤其是滑坡災(zāi)害,一旦發(fā)生降雨,一些地方就很容易發(fā)生滑坡,對(duì)人們的生命造成危害,對(duì)財(cái)產(chǎn)造成損失。因此,如何高效的對(duì)區(qū)域滑坡災(zāi)害進(jìn)行預(yù)防,是一個(gè)很重要的課題。數(shù)據(jù)挖掘是一種新興的數(shù)據(jù)分析技術(shù),它能通過分類方法進(jìn)行學(xué)習(xí),提取出規(guī)則,從而對(duì)未知的事物進(jìn)行預(yù)測。而滑坡災(zāi)害預(yù)測受多種因素影響,其中降雨等不確定因素存在難以獲取數(shù)據(jù)及有效處理等難題,為提高滑坡危險(xiǎn)性預(yù)測的準(zhǔn)確率,根據(jù)滑坡災(zāi)害發(fā)生相關(guān)理論及數(shù)據(jù)挖掘中決策樹分類原理,本文提出了兩種處理方法:基于不確定的C4.5決策樹算法和不確定模糊ID3決策樹算法,分別對(duì)滑坡危險(xiǎn)性進(jìn)行預(yù)測。本文首先進(jìn)行了滑坡災(zāi)害和數(shù)據(jù)挖掘相關(guān)的理論介紹,為后面的章節(jié)打下理論基礎(chǔ)。然后介紹了精度較高的不確定C4.5決策樹算法,并將它應(yīng)用于實(shí)例中,進(jìn)行效果檢驗(yàn)。接著在傳統(tǒng)模糊ID3算法的基礎(chǔ)上,利用積分思想,設(shè)計(jì)出不確定數(shù)據(jù)的模糊化處理方法,并結(jié)合模糊ID3算法,提出了一種新的不確定模糊ID3算法,建立不確定模糊ID3決策樹模型,對(duì)區(qū)域滑坡危險(xiǎn)性進(jìn)行預(yù)測,在簡化算法復(fù)雜度的基礎(chǔ)上,也保證了算法的精確度。最后,將傳統(tǒng)的算法和本文提出的算法進(jìn)行比較,分析兩種算法模型的優(yōu)缺點(diǎn),推斷出各自的適用場合,為以后該類問題的算法選擇提供參考。
[Abstract]:China is a country prone to geological disasters, especially landslide disasters. Once rainfall occurs, some places are prone to landslides, causing harm to people's lives and property losses. Therefore, how to efficiently prevent the regional landslide disaster is a very important subject. Data mining is a new technology of data analysis. It can learn from classification method, extract rules and predict unknown things. However, landslide disaster prediction is affected by many factors, such as rainfall and other uncertain factors, which are difficult to obtain data and deal with effectively, in order to improve the accuracy of landslide hazard prediction. According to the related theory of landslide occurrence and the classification principle of decision tree in data mining, this paper proposes two processing methods: C4.5 decision tree algorithm based on uncertainty and fuzzy ID3 decision tree algorithm based on uncertainty, respectively, to predict landslide risk. This paper firstly introduces the theory of landslide disaster and data mining, which lays a theoretical foundation for the later chapters. Then, the uncertain C4.5 decision tree algorithm with high precision is introduced, and it is applied to an example to verify the effect. On the basis of the traditional fuzzy ID3 algorithm, the fuzzy processing method of uncertain data is designed by using integral idea, and a new uncertain fuzzy ID3 algorithm is proposed based on fuzzy ID3 algorithm. An uncertain fuzzy ID3 decision tree model is established to predict the landslide risk in the region. Based on the simplification of the complexity of the algorithm, the accuracy of the algorithm is ensured. Finally, the traditional algorithm and the algorithm proposed in this paper are compared, the advantages and disadvantages of the two models are analyzed, and their respective applicable situations are inferred, which provides a reference for the selection of the algorithm for this kind of problem in the future.
【學(xué)位授予單位】：江西理工大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2015
【分類號(hào)】：P642.22;TP311.13

【參考文獻(xiàn)】

相關(guān)期刊論文 前3條

1 傅文杰;;GIS支持下基于支持向量機(jī)的滑坡危險(xiǎn)性評(píng)價(jià)[J];地理科學(xué);2008年06期

2 胡德勇;趙文吉;李小娟;李京;李家存;;不完備樣本條件下基于支持向量回歸模型的滑坡易發(fā)性評(píng)價(jià)[J];地理研究;2008年04期

3 姜琪文,許強(qiáng),何政偉;基于SVM多類分類的滑坡區(qū)域危險(xiǎn)性評(píng)價(jià)方法研究[J];地質(zhì)災(zāi)害與環(huán)境保護(hù);2005年03期

，

本文編號(hào)：1777492