Lu Han

Lu Han
Central University of Finance and Economics | CUFE · School of Management Science and Engineering

Doctor of Engineering

About

32
Publications
4,813
Reads
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
192
Citations
Citations since 2016
26 Research Items
164 Citations
2016201720182019202020212022010203040
2016201720182019202020212022010203040
2016201720182019202020212022010203040
2016201720182019202020212022010203040
Additional affiliations
September 2018 - September 2019
University of North Carolina at Chapel Hill
Position
  • Visiting Scholar
June 2014 - present
Central University of Finance and Economics
Position
  • Professor (Associate)
July 2012 - June 2014
Tsinghua University
Position
  • PostDoc Position
Education
September 2007 - July 2012
Beihang University (BUAA)
Field of study
  • Management Science and Engineering
September 2003 - July 2007
Beihang University (BUAA)
Field of study
  • Information Management and Information System

Publications

Publications (32)
Article
The most commonly used techniques for credit scoring is logistic regression, and more recent research has proposed that the support vector machine is a more effective method. However, both logistic regression and support vector machine suffers from curse of dimension. In this paper, we introduce a new way to address this problem which is defined as...
Article
Full-text available
In recent years, to improve predictive ability of corporate defaults has become an important problem. In this paper, regarding on characteristics of listed companies, we sampled 100 companies according to industry types, constructed wavelet structural model, experimented with wavelet decomposition proceeds to get low frequency and high frequency se...
Article
Full-text available
The advance of Internet technology has provided a convenient market platform for matching lending and borrowing parties, but many consumers still hesitate to use online borrowing. To better understand consumer behaviour in online borrowing, we use nationally representative survey data in China to explore factors affecting consumer use of one type o...
Article
Full-text available
The lack of standardized financial statements makes it difficult to determine the credit ratings of small and medium-sized enterprises (SMEs). Focusing on this problem, we construct an ensemble framework based on evidence theory. First, we change the sale amount to cash flow lift through a difference table. Then, we analyse consumer comments using...
Article
Full-text available
Aiming at the problems of subjective enhancement caused by the discretization of credit data and the lack of a multi-dimensional portrait of credit users in the current credit data research, this paper proposes an improved Fuzzy MLKNN multi-label learning algorithm based on MLKNN. On the one hand, the subjectivity of credit data after discretizatio...
Chapter
Fuzzy value is much important in credit risk management of commercial banks. However, there is not a perfect method of cluster analysis to analyze the data. The traditional clustering method is not suitable for fuzzy value clustering. Although lattice clustering uses lattice degree instead of Euclidean metric to calculate distance, it also has some...
Chapter
The research aims to establish a method of measuring the competitiveness of municipal commercial banks in Jing-Jin-Ji region and to give suggestions to these banks. Based on the existing research results, the research establishes measurement system using two-stage DEA method and combines the results with financial analysis. The research reaches thr...
Chapter
In this paper, we introduce a two-stage NER method to recognize the entities in online-sale comments. Firstly, we pick out the NN-labeled words from the pre-labeled dataset to generate the topic-based dictionaries, and in turn, these dictionaries are exploited to identify topic-related entities of each comment. Secondly, [−1, 1] windows are applied...
Conference Paper
Aiming at the problem of numerous variables in credit data, a large amount of sample data, and inability to intuitively reflect user portraits. This paper uses the MLKNN algorithm to perform multi-label learning on the credit data. According to the results of the algorithm training under 24 sets of k values, the optimal number of neighbor samples o...
Chapter
Full-text available
In this paper, we introduce a two-stage NER method to recognize the entities about papers occurring in MCM. Firstly, we pick out the NN-labeled and the NP-labeled words from the pre-labeled dataset to generate the preprocessed word frequency table and labeled the method-related words with “MET”. Secondly, [−3, 3] windows are applied to identify the...
Article
Full-text available
As has been demonstrated, the long short-term memory (LSTM) algorithm has the special ability to process sequenced data; however, LSTM suffers from high dimensionality, and its structure is too complex, leading to overfitting. In this research, we propose a new method, RFG-LSTM, which uses a rectified forgetting gate (RFG) to restructure the LSTM....
Article
Full-text available
Nowadays, credit is becoming more and more important and the progress of artificial intelligence technology brings new opportunities for the development of credit investigation. In this paper, we come up with a novel reasoning model for a credit investigation system based on the fuzzy Bayesian network which is a new convergence of technology and fi...
Article
Full-text available
One of the classical methods of clustering is kNN and it is a simple and widely used algorithm. But when it comes to the samples with unstructured and fuzzy values, it is not applicable enough course it is based on the Euclidean distance, which does not have practical significant for such data . But these kinds of data are much important in credit...
Article
Full-text available
Trends in capital markets are difficult to predict due to the complex linkages. Recent research has shifted from predicting single market trends to predicting multiple market co-movements. Based on the WTI oil price, Douglas stock index and Shenzhen composition index, we use wavelet transformation to extract the stable domains and identify the co-m...
Article
In this paper, we analyze the inherent evolutionary dynamics of financial and energy markets, study their interrelationships and carrying out predictive analysis tasks in an integrated nonparametric framework. We consider the daily closing prices of BRENT Index, Arca Natural Gas price, DJIA Index, and SZSE Index during January 2012 to January 2017...
Article
Ever increasing ordinal variables are being collected by the Personal Credit Reference System in China, however this system suffers from analysis of this kind of data, which cannot be calculated by Euclidean distance. In this study, we put forward a hybrid KNN algorithm based on Sugeno measure, and we prove that the error of this algorithm is small...
Article
How to classify users’ clusters, then analyze the characteristics of users, and finally make user personas, is one of the key issues to further explore the credit information system. In academic research, clustering algorithms are often implemented for user classification. The clustering algorithm is based on the sample distance measure. In the cre...
Chapter
Full-text available
Uncertainties included in soft classification all exist in classical mathematics. However, in daily life, much information has the extended fuzzy concept, and the clustering algorithm established on fuzzy set for fuzzy samples has been poorly investigated. To solve above problems, this study made improvements on DBSCAN and proposed a DBSCAN algorit...
Article
Full-text available
In this paper, we analyze the inherent evolutionary dynamics of financial and energy markets, study their interrelationships and carrying out predictive analysis tasks in an integrated nonparametric framework. We consider the daily closing prices of Brent Crude Oil, DJIA Index, Shenzhen Component Index (SZSE) from January 2012 to January 2017 from...
Article
Full-text available
In our increasingly connected business environment, an enterprise's financial downturn can spread to its economic associates, a phenomenon defined as default contagion. In academic research, the default contagion is usually measured by the default dependence degree. This paper utilizes the Copula metric method to capture the default dependence degr...
Article
Most recently, researchers have found support vector machine can provide better performance in the prediction of credit scoring. However, support vector machine is a black-box method and lacks rules for selecting good input variables. Similar to other artificial intelligence methods, they face the problem of “garbage in, garbage out”. Thus SVM is u...
Article
Credit risk management is one of the most important problems which commercial bank should face with. Normally, the optimal way of credit risk management is to forecast the default accurately before the loan. In the recent study, it is always using the structure model to forecast the default of the listed companies, because this model can realize th...
Article
Full-text available
Today the most commonly used techniques for credit scoring are artificial intelligence and statistics. In this paper, we started a new way to use these two kinds of models. Through logistic regression filters the variables with a high degree of correlation, artificial intelligence models reduce complexity and accelerate convergence, while these mod...
Article
Full-text available
The problem for studying and application of credit scoring models to appraisal of the loan to Chinese companies with uncertain linguistic information is the multiple attribute decision making (MADM) problems. In this paper, we investigate the multiple attribute decision making (MADM) problems for studying and application of credit scoring models to...
Article
Complex product R&D system has been developed from one-professional to a new pattern of distributed collaborative research and development. In the distributed collaborative R&D systems, modeling the flow of knowledge is very important and difficult. In this paper, the author gives the petri net-based distributed collaborative R&D system models, bas...

Network

Cited By

Projects

Projects (3)
Project
The alternative data of credit text has become an important supplement to the Personal Credit Report System (PCRS) of the People's Bank of China. However, due to its semantic fuzziness and inconsistent data structure, this kind of data cannot be directly integrated with current system. This project focuses on the research of credit evaluation methods of credit text alternative data, and intends to make a breakthrough in the following three aspects: 1) to develop the gold standard corpus of credit text data through two-dimensional entity recognition technology, so as to realize the automatic annotation of credit texts; 2) to design a multi-level Bert model, so as to realize the multi-label learning of annotated credit texts; 3) to explore the credit evaluation model of the labeled credit texts and current data in PCRS based on evidence fusion. In a word, this study can not only provide technical supports for the analysis of credit text alternative data, but also provide theoretical methods to expand data sources for PCRS.
Archived project
China Scholarship Council Award for Visiting Scholar (No. 201806495014)
Archived project
This project explores the two types of default risk measurement models recommended by the Basel committee, improves relevant methods, and verifies them with China's listed companies and personal credit data, in order to construct default risk measurement models which are suitable for the credit risk evaluation in China's commercial banks.