Big data technologies have a strong impact on different industries, starting from the last decade, which continues nowadays, with the tendency to become omnipresent. The financial sector, as most of the other sectors, concentrated their operating activities mostly on structured data investigation. However, with the support of big data technologies, information stored in diverse sources of semi-structured and unstructured data could be harvested. Recent research and practice indicate that such information can be interesting for the decision-making process. Questions about how and to what extent research on data mining in the financial sector has developed and which tools are used for these purposes remains largely unexplored. This study aims to answer three research questions: (i) What is the intellectual core of the field? (ii) Which techniques are used in the financial sector for textual mining, especially in the era of the Internet, big data, and social media? (iii) Which data sources are the most often used for text mining in the financial sector, and for which purposes? In order to answer these questions, a qualitative analysis of literature is carried out using a systematic literature review, citation and co-citation analysis.

Text Mining for Big Data Analysis in Financial Sector: A Literature Review

Peer to Peer (P2P) Lending Problems and Potential Solutions: A Systematic Literature Review

Most of the traditional Recommendation Systems (RSs) focus on recommending only the popular items as they deal with a single objective precision/popularity. However, focusing on the diversity of the items in the recommendation list is also equally important to improve its relevance to the user, i.e., it is required to view RSs as a multi-objective optimization problem. Nevertheless, owing to popularity and diversity to be conflicting with each other, it degrades the accuracy of the recommendation list. Therefore, in this work, we use a multi-objective optimization method to maintain a trade-off between the popularity and the diversity and obtain multiple trade-off solutions in a single run. We first incorporate Bhattacharyya Coefficient in an existing nonlinear similarity computation model to create a new similarity model named as Bhat_sim to increase the prediction accuracy of the exiting rating evaluation methods. Further, we formulate a multi-parent crossover mechanism NewCross in the proposed multi-objective recommendation filtering NewCrossPMOEA which preserves the order and the frequency in the parents genes to bring good objectivity in the trade-off of recommending popular and diverse items in the recommendation list. The obtained results on the Movielens dataset demonstrate that the NewCrossPMOEA performs superior in terms of average precision, diversity, and novelty to its competing methods. Moreover, the Pareto-dominance concept of NewCrossPMOEA suggests multiple recommendation solutions of diverse and novel items to the target users in a single run.

Multi-objective item evaluation for diverse as well as novel item recommendations

At the early of 21st century, In the UK Peer2Peer (P2P) lending began with Zopa at the start of 2005. After that, it has grown rapidly in United States, China and some other countries. The main challenge for individual investors in the P2P lending market is to allocate their money efficiently through different loans by accurately evaluating the credit of each loan. The traditional ranking models cannot conform to the needs of individual investors in P2P lending because they do not provide natural mechanisms for asset allocation. P2P loans do not have the presence of traditional financial institutions. In this study, we propose a novel method for analyzing data for this emerging market. We have designed a credit risk model based on advanced machine learning methods, capable of assessing the returns and risks of individual loans. We also applied a selection feature method to eliminate the irrelevant features for improving the efficient of machine learning models. We conducted experiments on real-world data sets from P2P lending markets. The experimental results show that the proposed model can improve the efficiency of investment compared to the existing methods of P2P lending.

Improving Credit Risk Prediction in Online Peer-to-Peer (P2P) Lending Using Feature selection with Deep learning

We experimentally investigate a new variant of the trust/investment game that captures some key features of internet peer-to-peer (P2P) lending: the borrower specifies the amount of money required and makes a contingent promise about the value of the generally higher repayment prior to the investor’s decision to lend the required sum or not. We examine the role played by two factors related to traditional Chinese culture and ethics: whether (i) relational closeness between the actors and (ii) the ability of the actors to observe each other’s identity after the repayment decision (identifiability) affect the borrowers’ decisions to make the promised repayments and ultimately the consequent aggregate realized social benefits. Using a two-by-two factorial design, we conduct four treatments in China where these factors are hypothesized to be especially salient, and also perform the identifiability treatment in New Zealand as a cultural control. We find that in China both manipulations are positively and significantly related to the probability of a repayment promise being kept. Moreover, these two factors are substitutes for each other. In New Zealand, there was no significant identifiability effect on promise keeping. The effectiveness of identifiability in China but not in New Zealand resulted in a significantly higher proportion of promises being kept when agents were identifiable in China than in New Zealand. Over time, relational closeness and identifiability both led investors in China to accept more proposals, resulting in more investment and the creation of greater social surplus.

Promise Keeping, Relational Closeness, And Identifiability: An Experimental Investigation In China

Online social lending has facilitated the ability of borrowers to reach lenders for financing support. With the increasing number of social lending projects, it is becoming very difficult for lenders to find appropriate projects to invest in, and for borrowers to get the funds they need. Project recommendation techniques provide a promising way to solve this problem to some degree, by recommending borrowers’ projects to lenders who are able to invest. Unfortunately, current loan project recommendations only explore some structured information to match borrowers and lenders, so they cannot achieve a satisfactory way to solve the problem very well. In this study, we innovatively mine a huge amount of unstructured data, the text data of borrowers’ and lenders’ motivations, to provide loan project recommendations that solve the problem of mismatches between borrowers and lenders. We present a motivation-based recommendation approach that uses text mining and classifier techniques to identify borrowers’ and lenders’ motivations. Using a dataset from the well-known social lending platform Kiva, our experiment results show that, compared with prior works, the proposed approach improves project recommendations in inactive lender groups and unpopular loan groups, which shows the superiority of the proposed approach in addressing data sparsity and cold start problems in loan project recommendations. This study thus initiates an attempt to solve the information overload problem and improve matching between borrowers and lenders through mining big unstructured text data found in a large number of P2P platforms.

Hong Zhu

Papers

Mining Social Lending Motivations for Loan Project Recommendations