【摘要】：改革開放以來，中小企業(yè)在國民經(jīng)濟(jì)和社會(huì)發(fā)展中發(fā)揮著巨大作用，近年來浮現(xiàn)的中小企業(yè)融資難問題已成為社會(huì)各界關(guān)注的焦點(diǎn)。一方面，由于財(cái)務(wù)制度不完善、抵御風(fēng)險(xiǎn)能力弱、銀企信息不對(duì)稱等原因，中小企業(yè)很難從銀行貸到款，融資問題已成為阻礙其發(fā)展的關(guān)鍵因素；另一方面，銀行業(yè)競爭日益激烈，在目標(biāo)客戶選擇上的趨同性致使信貸資源向大客戶集中，銀行議價(jià)能力減弱，利差空間縮小，急需開辟新的業(yè)務(wù)領(lǐng)域，制造新的利潤增長點(diǎn)。如何開發(fā)出合適的中小企業(yè)貸款決策模型，以實(shí)現(xiàn)風(fēng)險(xiǎn)、收益的平衡與中小企業(yè)和銀行的共贏，成為擺在我們面前的重要課題。 本文在對(duì)中小企業(yè)融資現(xiàn)狀和風(fēng)險(xiǎn)特征進(jìn)行充分調(diào)研與分析的基礎(chǔ)上，通過研究銀行貸款決策的國內(nèi)外文獻(xiàn)和相關(guān)技術(shù)，建立了中小企業(yè)銀行貸款決策模型。該模型計(jì)算量小，應(yīng)用性強(qiáng)，，預(yù)測較準(zhǔn)確，能有效運(yùn)用于銀行對(duì)中小企業(yè)的貸款決策。 本文的研究工作主要有以下幾個(gè)方面： 1.中小企業(yè)財(cái)務(wù)評(píng)價(jià)指標(biāo)體系的建立。本論文運(yùn)用財(cái)務(wù)分析方法對(duì)中小企業(yè)財(cái)務(wù)評(píng)價(jià)指標(biāo)進(jìn)行初選，運(yùn)用計(jì)量經(jīng)濟(jì)軟件EViews建立回歸模型，經(jīng)模型檢驗(yàn)后最終建立由8個(gè)核心指標(biāo)構(gòu)成的中小企業(yè)財(cái)務(wù)評(píng)價(jià)指標(biāo)體系，從而確定出影響銀行對(duì)中小企業(yè)貸款的決定因素。 2.采用組合算法建立中小企業(yè)銀行貸款決策模型。本論文采集上市公司相關(guān)財(cái)務(wù)數(shù)據(jù)作為實(shí)驗(yàn)樣本并進(jìn)行標(biāo)準(zhǔn)化處理，以MATLAB作為開發(fā)語言，通過遺傳算法改進(jìn)RBF神經(jīng)網(wǎng)絡(luò)的參數(shù)，用梯度下降法對(duì)結(jié)果進(jìn)一步尋優(yōu)，進(jìn)而得出預(yù)測結(jié)果。 3.實(shí)驗(yàn)仿真與模型應(yīng)用。本文利用經(jīng)訓(xùn)練的中小企業(yè)銀行貸款決策模型，分別使用經(jīng)遺傳算法優(yōu)化和未經(jīng)過優(yōu)化的RBF神經(jīng)網(wǎng)絡(luò)進(jìn)行仿真實(shí)驗(yàn)，結(jié)果顯示引入遺傳算法提高了模型預(yù)測的準(zhǔn)確率。進(jìn)一步將該模型應(yīng)用于某商業(yè)銀行實(shí)際的中小企業(yè)貸款中，通過對(duì)預(yù)測結(jié)果的分析為貸款決策提供依據(jù)。
[Abstract]:Since the reform and opening up, small and medium-sized enterprises (SMEs) have played a great role in the national economy and social development. In recent years, the financing difficulties of SMEs have become the focus of attention. On the one hand, because the financial system is not perfect, the ability to resist risks is weak, and the information of banks and enterprises is not symmetrical, it is very difficult for small and medium-sized enterprises to get loans from banks, so the financing problem has become the key factor that hinders their development. On the other hand, the competition of the banking industry is becoming more and more intense, and the convergence in the choice of target customers leads to the concentration of credit resources to the large customers, the weakening of the bargaining power of the banks, the narrowing of the margin space, the urgent need to open up new business fields and the creation of new profit growth points. How to develop a suitable loan decision model for SMEs to realize the balance of risk and income and the win-win situation between SMEs and banks has become an important subject in front of us. Based on the investigation and analysis of the current financing situation and risk characteristics of small and medium-sized enterprises (SMEs), this paper establishes a decision model of bank loans for small and medium-sized enterprises (SMEs) by studying the domestic and foreign literature and related technologies of bank loan decision-making. The model has the advantages of small calculation, strong application, accurate prediction, and can be effectively applied to the loan decision of the bank to the small and medium-sized enterprises. The main research work of this paper is as follows: 1. The establishment of financial evaluation index system for small and medium-sized enterprises. In this paper, the financial evaluation index of small and medium-sized enterprises is selected by the method of financial analysis, and the regression model is established by using the econometric software EViews, and the financial evaluation index system of small and medium-sized enterprises which is composed of eight core indicators is finally established after the model test. In order to determine the impact of bank loans to small and medium-sized enterprises to determine the factors. 2. The combination algorithm is used to establish the decision model of bank loan for small and medium-sized enterprises. In this paper, the relevant financial data of listed companies are collected as experimental samples and standardized processing. With MATLAB as the development language, the parameters of RBF neural network are improved by genetic algorithm, and the results are further optimized by gradient descent method. Then the prediction results are obtained. 3. Experimental simulation and model application. In this paper, the trained SME bank loan decision model is used to simulate the RBF neural network, which is optimized by genetic algorithm and unoptimized, respectively. The results show that the accuracy of model prediction is improved by introducing genetic algorithm. Furthermore, the model is applied to a commercial bank to provide the basis for the loan decision through the analysis of the forecast results.
【學(xué)位授予單位】：上海交通大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2014
【分類號(hào)】：F276.3;F832.4;TP183

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 張程;李文雯;張振新;;公司規(guī)模與資產(chǎn)負(fù)債率:來自上市公司的證據(jù)[J];財(cái)經(jīng)問題研究;2010年11期

2 何自力;;銀行中小企業(yè)貸款的效益與風(fēng)險(xiǎn)分析[J];金融論壇;2006年01期

3 張超群;鄭建國;錢潔;;遺傳算法編碼方案比較[J];計(jì)算機(jī)應(yīng)用研究;2011年03期

4 柯孔林;馮宗憲;;基于粗糙集和神經(jīng)網(wǎng)絡(luò)集成的貸款風(fēng)險(xiǎn)5級(jí)分類[J];控制理論與應(yīng)用;2008年04期

5 李朝靜;唐幼純;黃霞;;BP神經(jīng)網(wǎng)絡(luò)的應(yīng)用綜述[J];勞動(dòng)保障世界(理論版);2012年08期

6 張友棠,楊軼;基于“5C”的企業(yè)信用評(píng)價(jià)標(biāo)準(zhǔn)化體系的構(gòu)建[J];世界標(biāo)準(zhǔn)化與質(zhì)量管理;2005年09期

7 陳曉紅;張澤京;王傅強(qiáng);;基于KMV模型的我國中小上市公司信用風(fēng)險(xiǎn)研究[J];數(shù)理統(tǒng)計(jì)與管理;2008年01期

8 田永強(qiáng);層次分析法在銀行貸款投放決策中的應(yīng)用[J];統(tǒng)計(jì)與決策;2004年07期

9 鄧?yán)^雄;李志舜;梁紅;;確定RBF神經(jīng)網(wǎng)絡(luò)參數(shù)的新方法[J];微處理機(jī);2006年04期

10 郭戰(zhàn)琴,周宗放;基于VaR約束的商業(yè)銀行貸款組合多目標(biāo)決策[J];系統(tǒng)工程理論方法應(yīng)用;2005年02期

本文編號(hào)：2321895