事件式精准营销在哈尔滨银行小企业信贷业务应用研究

发布时间：2019-03-21 15:04

【摘要】：2014年宏观经济增速放缓，服务主导型经济正在逐渐取代工业主导型经济。2014年我国货币乘数居高位，意味着各家商业银行依然需要面临较高水平的法定存款准备金率，银行面临较大的资金压力。《中国人民银行关于进一步推进利率市场化改革的通知》一文下发，中国人民银行全面放开了金融机构贷款利率水平，商业银行的存款和贷款竞争将达到一个新的白热化阶段，银行的利润水平将会大幅下降，在资金和技术方面处于弱势的城市商业银行将面临巨大挑战。哈尔滨银行是东北地区最大的上市城市商业银行，采用事件式精准营销将对哈尔滨银行提高营销效率获得长远发展具有重要意义。事件式精准营销在哈尔滨银行小企业信贷业务的应用是一个交叉学科的研究，涉及计算机科学、营销管理学、经济学多个学科，使用了访谈、问卷、专家打分、历史文献分析等方法进行定性研究，也使用了C5.0决策树算法模型和boostings数据推进等实证方法进行定量研究。对哈尔滨银行事件式精准营销的实施背景的分析发现，，哈尔滨银行在发展过程中具有很好的数据库架构和数据质量，小企业信贷经验丰富、产品多样、营销渠道成熟，具备进行事件式精准营销的条件。经过访谈调研设计了精准营销事件库，通过专家打分和数据库支持检测发现，贷款到期是事件库中最具有试点价值的事件。针对贷款到期事件设计的被解释变量为续贷行为是否发生，解释变量包括客户基本信息变量、客户存款变量、客户贷款变量，因为研究的实际意义在于根据模型产生的续贷名单进行精准营销，所以更加倾向于续贷的建模和预测。对贷款到期客户解释变量和被解释变量进行数据探查发现，只有部分数据能够从数据库中取到数据，而且具有离群值和缺失值的问题数据存在。抛弃和均值替换之后的样本数据可以进行数据建模，建模过程中出现使用boostings数据推进和样本集分割等统计学方法进行模型优化，训练样本集和测试样本集的预测精度接近95%，将模型应用于预测全新的数据，预测精度接近了90%，这说明模型可以用于预测贷款到期客户有续贷可能性的名单，极大缩小了营销范围，降低了营销成本，提高了营销效率。
[Abstract]:The macro-economic growth rate slowed in 2014, and the service-led economy is gradually replacing the industrial-led economy. In 2014, China's currency multiplier is at a high level, which means that commercial banks still need to face a higher level of Statutory deposit reserve ratio. The Bank of China is facing considerable capital pressure. The people's Bank of China issued the Circular on further promoting the Reform of interest rate marketization. The people's Bank of China has fully liberalized the interest rate level of financial institutions. The deposit and loan competition of commercial banks will reach a new hot stage, the profit level of banks will decrease greatly, and the city commercial banks which are weak in capital and technology will face great challenges. Harbin Bank is the largest listed city commercial bank in Northeast China. It is of great significance for Harbin Bank to improve marketing efficiency and obtain long-term development by adopting event-based accurate marketing. The application of event-based accurate marketing in the credit business of small enterprises of Harbin Bank is a cross-disciplinary study involving computer science, marketing management and economics. It uses interviews, questionnaires, and expert scores. Qualitative research is carried out by historical literature analysis and quantitative research is carried out by using C5.0 decision tree algorithm model and boostings data advancement method. Based on the analysis of the implementation background of the event-based accurate marketing of Harbin Bank, it is found that Harbin Bank has a good database structure and data quality in the process of development, the credit experience of small enterprises is rich, the products are diverse, and the marketing channels are mature. Has the condition to carry on the event type accurate marketing condition. A precise marketing event database is designed through interview and investigation. It is found that loan maturity is the most valuable event in the event bank through expert scoring and database support testing. The interpreted variables designed for loan maturity events are whether the renewal behavior occurs or not. The explanatory variables include customer basic information variable, customer deposit variable, customer loan variable, and so on. Because the practical significance of the research lies in the precise marketing according to the renewal list generated by the model, it is more inclined to model and forecast the renewal. It is found that only part of the data can be retrieved from the database and the problem data with outliers and missing values exists through the data exploration of the interpreted variables and interpreted variables of the loan maturity customers. The sample data after abandonment and mean replacement can be used for data modeling. In the modeling process, statistical methods such as boostings data advancement and sample set segmentation are used to optimize the model, and the prediction accuracy of training sample set and test sample set is close to 95%. Applying the model to the new forecast data, the prediction precision is close to 90%. This shows that the model can be used to predict the list of loan renewal possibility of customers due to loan, which greatly reduces the marketing scope, reduces the marketing cost, and improves the marketing efficiency.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F832.4

【相似文献】