Building an analytics framework for predicting customer default****

这项研究旨在提供数据分析和预测建模方面的专业知识，以帮助丰业银行准备一个适合分析的数据集，并开发和实施客户违约预测模型。该项目将帮助完成以下任务：1。从银行确定的不同系统中提取、组织和整合数据; 2.修复数据质量问题（不一致、缺失值、重复等），3.将原始数据转换为衍生变量，这些变量可以是有用的模型输入并执行特性分析。使用机器学习算法开发预测模型。** 随着公司每天积累大量数据，他们必须找到一种更有效的方法来管理数据，并主动评估风险，以帮助减少甚至防止客户违约。这项研究将帮助该公司为其平台配备新的功能，以有效地处理各种数据，并利用这些数据来预测客户违约。此外，该项目将为高素质人员（HQP）提供宝贵的学习和培训环境，他们将开发为金融机构开发预测建模技术的技能和专业知识。

This research aims to deliver expertise in data analytics and predictive modeling to help Scotiabank prepare a dataset that is amenable for analysis and to develop and implement predictive models for customer default. The project will help accomplish the following tasks; 1. Extract, organize and integrate data from the different systems identified by the bank, 2. Fix data quality issues (inconsistencies, missing values, duplication, etc.), 3. Transform the raw data into derived variables which can be useful model inputs and perform characteristics analysis, 4. Develop predictive models using machine learning algorithms. **With large volumes of data accumulated by companies every day, it is imperative that they find a more efficient and effective way to manage their data, and proactively assess risk to ideally help reduce and even prevent customer default. This research will help the company equip its platform with new capabilities to efficiently handle diverse data, and use it to predict customer default. Additionally, the project will provide a valuable learning and training environment for Highly Qualified Personnel (HQP) who will develop the skills and expertise to develop predictive modeling techniques for financial institutions.**********