An Enterprise Credit Management Information System Design

With the prosperous development of financial globalization and financial market, credit risk management was paid particular attention to in financial industry. In addition, the continuous expansion of large database or data center in scale sharply increased data source and data item became richer. On the other hand, data collection was more frequent and data growth more swift and violent. The adding of two factors let various areas from government to commercial banks urgently need one unified and effective enterprise credit evaluation system.Targeting at the above problem, this paper designed credit management system. The overall design on management system took the status of credit management at present as analysis basis and government database to construct public financial data platform, so as to help commercial banks to achieve information reform and effectively establish the financial market environment of data cleaning on enterprise information rating, and finish efficient design. In addition, it integrated the current economic system, enabling it to adapt to the characteristics of credit risk management system. As per the daily transaction process and operation process of credit business, it improved information management function of credit risk management platform. This paper proposed ternary database overall structure, i.e. the data processing mode of enterprise information database, principal database and credit risk database, and how they conduct collaborative operation. Later on, it focused on expounding on how to allocate the final credit risk database, and designed the basic operation process.The core of this article was to resolve on how to clean data source and let it become effective data to support enterprise information analysis. In this design, this paper firstly analyzed the quality problems such as heterogeneous data source, diversified butt joint mode, loose rules, data repetition, and inaccuracy existing in data source. Data cleaning rule and working procedure was designed as per procedure, and the quantity of data source was confirmed. As per the data analysis results of “dirty data”, appropriate detecting algorithm, explicit strategy and evaluation method were selected to implement data conversion and cleaning procedure. Data cleaning and conversion patterns were further resolved, the query and matching language were designated as far as possible, and no conversion code automatic generation was resolved, thereby utilizing these data to construct available bank credit risk model.The innovative point of this article lies in that it proposed core word-based repeating data detection model. Basing on such model, it developed a set of repeating data detection procedure and extracted data to conduct verification analysis. It analyzed the composition characteristics of enterprise name in enterprise credit basic database, and constructed characteristic attribute chart of enterprise name. It innovatively utilized “atom segmentation” method in design, and constructed industry dictionary chart, regional dictionary chart, suffix dictionary chart and filler term dictionary chart. Then it developed characteristic attribute extraction procedure and similarity calculation procedure. The final part of this paper conducted duplicate detection on two batches of experimental data and presented results analysis. The repeating data detection model proposed has practicality and expandability, which is not only effective for duplicate detection of enterprise name, but also suitable for the duplicate detection of general record. It has important practical significance in resolving repeating data and low department data correlation rate in enterprise credit basic database.