Real Transaction Data Research Articles

The effects of macroprudential policies on house prices: Evidence from non-ordinary house purchasing policy using China real transaction data

This paper contributes to the international policy debate on the effect of macroprudential policy on housing prices. Our identification relies on the fact that ordinary housing and non-ordinary housing in Beijing face different loan-to-value (LTV) restrictions and transaction tax rates. This paper uses regression discontinuity (RD) design to test whether the differentiated regulation of ordinary housing and non-ordinary housing caused a discontinuous change in housing prices with house-level transaction data. We find that housing prices have a significant discontinuous decline at the cutoff point caused by macroprudential policy, which we take 140 square meters since it is the identification standard of non-ordinary housing. Further, we carry out RD-DID analysis to investigate the influence of loan-to-value (LTV) restrictions on housing prices. We find that around 2% decline in house prices is caused by credit policy.

A proposed model for enhancing e-bank transactions: an experimental comparative study

In this paper, we introduce a novel approach to address the dynamic prediction of customer activity in electronic payment transactions for individual clients. Our approach is founded on customer online payment transaction records from registered UK-based online retailers between 01/12/2009 and 09/12/2011. These retailers primarily specialize in unique gift items for various occasions, catering to a wide range of clients, including wholesalers. We used classification analysis based on the correlation coefficient to measure and describe a customer's electronic payment capability based on the quality of products they purchase. Furthermore, we trained multi-layered models (linear model, deep learning, random forest, and support vector machines (SVM)) to capture the dynamics of e-bank transaction reinforcement for retail customers using machine learning. Real transaction data from a UK online retailer was employed in our study. The experimental results consistently demonstrated the effectiveness of our proposed strategy.

Mobile Food Ordering Apps, Restaurant Performance, and Customer Satisfaction

Mobile food ordering apps (MFOAs) have become increasingly prevalent in the food service industry. This study investigated the links between intensity of MFOA use, restaurant performance, and customer satisfaction in dine-in restaurants using real transaction data and crowdsourced reviews. Our distinctive datasets stem from a renowned chain of restaurants in China as well as from Da Zhong Dian Ping (referred to as the Chinese equivalent of Yelp), spanning the years 2018 to 2021. Using a high-dimensional fixed effects model, we found that the intensive use of MFOA is positively correlated with an increase in restaurant dine-in revenue through higher table turnovers. However, MFOA is negatively correlated with dine-in restaurant consumer satisfaction, which was associated with lack of adequate, favorable, and high-quality comments. We also conducted several heterogeneity analyses to investigate how selected restaurant characteristics may moderate these effects. Thus, our study provides valuable theoretical insights and practical implications for academia and food service practitioners.

Research on Financial Fraud Detection Models Integrating Multiple Relational Graphs

The current fraud risk in digital finance is increasing year by year, and the mainstream solutions rely on the inherent characteristics of users, which makes it difficult to explain fraud behaviors and fraud behavior patterns are less researched. To address these problems, we propose an integrated multiple relational graphs fraud detection model Tri-RGCN-XGBoost, which analyzes the impact of user association patterns on fraud detection by mining the behavioral associations of users. The model builds a heterogeneous information network based on real transaction data, abstracts three types of bipartite graphs (user–device, user–merchant, and user–address), aggregates the information of the user’s neighbor nodes under the three types of behavioral patterns, and integrates the graph convolution classification results under the three behavioral patterns with the XGBoost model to achieve fraudulent user detection with integrated multiple relational graphs. The results show that the performance of this model in fraud identification is significantly improved, especially in reducing the fraudulent user underreporting rate. Further, the behavioral associations that play a key role in fraud user identification are analyzed in conjunction with shape value to provide a reference for fraud pattern mining.

AWI-BS: An adaptive weight incentive for blockchain sharding

The sharding technique enables blockchain to process transactions in parallel by dividing blockchain nodes into small groups, each of which handles a subset of all transactions. One of the issues with blockchain sharding is generating a large number of cross-shard transactions that need to be checked on the output shard as well as the destination shard. Our analysis suggests that the processing efficiency of cross-shard transactions is consistent with the barrel effect, i.e., that efficiency is more dependent on slower processing shard. Most of the existing studies focus on how to deal with cross-shard transactions, but neglecting the fact that the relative independence between sharding results in different incentive costs between sharding. We perform a sharding analysis on 100,000 real transactions data on Ethereum, and the results show that there is a large difference in gas prices between different shards indeed. In this paper, we propose an Adaptive Weight Incentive (AWI) for Blockchain Sharding, which uses adaptive weight in place of traditional incentive, to address the problem of differing incentive costs for each shard. Take Ethereum as an example, AWI-BS computes the weight of a transaction as a function of a combination of the underlying gas price, the latency of the transaction, and the urgency of the transaction. Then the node chooses which transaction to pack based on the AWI-BS. Lastly, we also perform an in-depth analysis of AWI-BS's security and effectiveness. The evaluation indicates that AWI-BS outperforms the other alternatives in terms of transaction confirmation latency, transaction hit rate, and system throughput.

Feature generation and contribution comparison for electronic fraud detection

Modern money transfer services are convenient, attracting fraudulent actors to run scams in which victims are deceived into transferring funds to fraudulent accounts. Machine learning models are broadly applied due to the poor fraud detection performance of traditional rule-based approaches. Learning directly from raw transaction data is impractical due to its high-dimensional nature; most studies construct features instead by extracting patterns from raw transaction data. Past literature categorizes these features into recency, frequency, monetary, and anomaly detection features. We use various machine learning algorithms to examine the performance of features in these four categories with real transaction data; we compare them with the performance of our feature generation guideline based on the statistical perspectives and characteristics of (non)-fraudulent accounts. The results show that except for the monetary category, other feature categories used in the literature perform poorly regardless of which machine learning algorithm is used; anomaly detection features perform the worst. We find that even statistical features generated based on financial knowledge yield limited performance on a real transaction dataset. Our atypical detection characteristic of normal accounts improves the ability to distinguish them from fraudulent accounts and hence improves the overall detection results, outperforming other existent methods.

Seekers’ complaint behavior in crowdsourcing: An uncertainty perspective

This research aims to understand how task descriptions on crowdsourcing platforms influence seekers’ complaint behavior, which has received little if any attention in research. From an uncertainty perspective, we develop a research model that depicts how seekers’ task description (task description uncertainty), solvers’ expertise (solver uncertainty), and task duration jointly influence seekers’ complaint behavior. The model was tested with real transaction data from ZBJ.com, a popular crowdsourcing platform in China. The results indicate that task description precision can reduce seekers’ complaint behavior while task description completeness counter-intuitively increases complaint behavior. Moreover, solvers’ expertise decreases seekers’ complaint behavior and at the same time attenuates the effects of task description on seekers’ complaint behavior. In addition, task duration negatively moderates the effect of task description completeness on complaint behavior.

How do e-stores grow their market share?

PurposeThe purpose of this paper is to identify how fashion and cosmetics e-stores compete and grow to help e-commerce managers set the corresponding marketing strategy. It describes the relevance of customer acquisition and retention to market share as the essential performance metrics.Design/methodology/approachAn empirical generalization approach where patterns in data appearing across studies are described by a mathematical or graphical method is used. To do that, the authors observed real transactional data and the effect of how e-stores benefit from new and returning customers and gain a larger market share. The authors have analysed behavioural data from nearly 124,000 e-commerce customers in two highly popular product categories (fashion and cosmetics) in the size of 10,000,000 euros in sales or more.FindingsFashion and cosmetics e-stores with more market penetration tend to have a higher market share measured both by the number of total purchases and the number of sales in euro. In other words, market penetration is a solid predictor of market share in all circumstances. Interestingly, no significant difference in loyalty has been observed in relation to market share growth except in the situation where the market partition was excluded from the product category.Research limitations/implicationsThe businesses under study derived only from one country and only two product categories were observed. Thus, there is a potential limitation in generalizing the findings to the whole e-commerce market from a geographical and category perspective. The length of the observation period may also play a role as a longer period increases the chance of repeat buying.Practical implicationsE-commerce managers can gain long-term market share growth mainly via higher market penetration (acquisition of new customers) and should avoid misleading overfocus on loyalty tactics (retention of current customers). The study also provides important benchmarks for e-commerce businesses in the fashion and cosmetic categories.Originality/valueIn the market share growth literature, only a handful of studies focus on stores and not on products. Moreover, there is a dominance of fast-moving consumer goods categories. Surprisingly, studies analysing ever-growing e-commerce markets are scarce. Thus, this research is original because it describes, using empirical data, how brands online, at the store level and within the fashion and cosmetics category, grow their market share. It is also one of the few studies that work with real business transactional data.

Decision Forest: A Nonparametric Approach to Modeling Irrational Choice

Customer behavior is often assumed to follow weak rationality, which implies that adding a product to an assortment will not increase the choice probability of another product in that assortment. However, an increasing amount of research has revealed that customers are not necessarily rational when making decisions. In this paper, we propose a new nonparametric choice model that relaxes this assumption and can model a wider range of customer behavior, such as decoy effects between products. In this model, each customer type is associated with a binary decision tree, which represents a decision process for making a purchase based on checking for the existence of specific products in the assortment. Together with a probability distribution over customer types, we show that the resulting model—a decision forest—is able to represent any customer choice model, including models that are inconsistent with weak rationality. We theoretically characterize the depth of the forest needed to fit a data set of historical assortments and prove that with high probability, a forest whose depth scales logarithmically in the number of assortments is sufficient to fit most data sets. We also propose two practical algorithms—one based on column generation and one based on random sampling—for estimating such models from data. Using synthetic data and real transaction data exhibiting nonrational behavior, we show that the model outperforms both rational and nonrational benchmark models in out-of-sample predictive ability. This paper was accepted by Chung Piaw Teo, optimization.

Order processing task allocation and scheduling for E-order fulfilment

This paper mainly studies a task allocation and scheduling problem in the multi-thread fulfilment process of electronic order, which seeks to minimise the makespan under thread constraints and order precedence constraints. The problem is formulated as a Mixed Integer Programming (MIP) model and a novel depth-first heuristic is proposed to solve it. The depth-first heuristic shows high effectiveness and efficiency, compared with the current policy and the genetic algorithm in both small/medium-scale and large-scale cases from the real transaction data. In addition, two extensions on precedence constraint reduction and resource allocation are discussed to further improve and manage the e-order fulfilment process.

Recommendation system with minimized transaction data

This paper deals with the recommendation system in the so-called user-centric payment environment where users, i.e., the payers, can make payments without providing self-information to merchants. This service maintains only the minimum purchase information such as the purchased product names, the time of purchase, the place of purchase for possible refunds or cancellations of purchases. This study aims to develop AI-based recommendation system by utilizing the minimum transaction data generated by the user-centric payment service. First, we developed a matrix-based extrapolative collaborative filtering algorithm based on open transaction data. The recommendation methodology was verified with the real transaction data. Based on the experimental results, we confirmed that the recommendation performance is satisfactory only with the minimum purchase information.

Economic spillover from renewable energy industries: an input-output analysis

ABSTRACT This study aims to identify the changes in the economic relations between industries resulting from the dissemination of renewable energy in Korea. Although previous studies have examined the ripple effect of renewable energy using an input-output table, there are limited studies that have performed an accurate analysis due to the absence of real economic transaction data for renewable energy and other industries. Thus, an input-output analysis was performed using actual data for 2010 and 2015 as well as 2020 estimated by RAS method. Backward and forward linkage effects and inducement effects were estimated. Specifically, the renewable energy value chain industries were classified into Raw material, Machine part, Infra structure, and End product. The main findings are: (1) the effect ratio of renewable energy industry decreased to 0.9572 reaching a similar level to the existing electric power industry; (2) the production inducement effects were 1.1278(2010), 0.7016(2015), and 0.7203(2030), and value-added inducement effects were 0.6043(2010), 0.3860(2015), and 0.4069(2020); and (3) the proportion of inducement effect on the service industry decreased approximately from 40% to 34%. These findings show how the renewable energy industry has changed from the early stages to the present and its economic importance in other industries resulting from renewable energy production and distribution. This study provides basic data for diagnosing the economic spillover and establishing energy policies due to the dissemination of renewable energy.

Forecasting Sales in the Supply Chain Based on the LSTM Network: The Case of Furniture Industry

Purpose: The aim of the article is to develop an algorithm for forecasting sales in the supply chain based on the LSTM network using historical sales data of a furniture industry company. Design/Methodology/Approach: Machine learning was used to analyze the data. The method of predicting the behavior of sales value in a specific time horizon in terms of a time series was presented. The LSTM network was used to predict sales. The network used is a special case of recursive neural networks with an important difference in the repeating module. Due to the fact that the activities are carried out on time series, the data was analyzed in terms of the stationarity of such series or trends and seasonal effects. The data used in the analysis includes the daily sales values of a group of certain furniture collections over a specified time horizon. The stationarity of the time series can have a significant impact on its properties and behavior prediction, where failure to bring the time series to the correct form of stationarity can lead to false results. Findings: The result of the research was the analysis of sales forecasting in the supply chain based on machine learning. As a result of the data transformations, the algorithm was able to recognize and learn long-term relationships. Practical Implications: The presented method of predicting the behavior of sales value in a specific time horizon allows for a look at the forecasting of demand in terms of the supply chain. The sales data of a company from the furniture industry were used for the analysis. Originality/Value: A novelty is the use of the LSTM network trained on real transaction data of a furniture company that has based its business on the supply chain and cooperates with its suppliers and recipients in Central and Eastern Europe.

Latent association graph inference for binary transaction data

A novel approach to the problem of statistical inference for multivariate binary transaction data is proposed. A fundamental question that arises from this data, often referred to as market basket data, is how the items relate to one another. These relationships are naturally expressed by a graph and transactions can be modeled as samples of cliques from this association graph. A hierarchical model is developed that follows from this generative idea, along with an MCMC sampling procedure that handles large datasets and allows inference on a broad set of parameters. This model provides a sparser representation of associations between items as compared with frequent itemset mining (FIM) output, without sacrificing predictive accuracy. Additionally, by allowing inference on a broad set of parameters, the model provides a deeper level of insight into transaction data. Empirical results are provided on applications of this model to simulated data and real transaction data from Instacart.

Impact of electric vehicles charging demand on distribution transformers in an office area and determination of flexibility potential

Electric vehicles (EVs) have a high potential in reducing greenhouse gas emissions and are able to achieve other advantages such as a reduction in local air pollution and increasing energy security. As a result, EVs are rapidly increasing in popularity, electrifying the transportation sector. This poses a serious problem for the grid as existing distribution grids were mainly sized in the pre-EV era. In this paper, a method is proposed to determine the charging demand of future EV fleets in an office area and determine its flexibility potential. Office charging is studied as it differs from residential charging and a limited number of studies focused solely on office charging. The study is an empirical study and is based on analysing real transaction data of 42 EVs charging for over a year at Utrecht Science Park, Utrecht, the Netherlands, the considered case study. The transaction data allows for an examination of the impacts of future EV charging demand in an office area. The results based on a future scenario study show that in 2050, 4 out of 7 studied transformers are overloaded. This is followed by an analysis on the mitigation of the determined impact. This analysis also determined the flexibility in EV demand, around 50% of the EV demand can be delayed for more than 8 h. When this flexibility is used, overloading of 3 out of 4 transformers could be mitigated. This paper shows that placement of charging stations should be strategically performed, considering the capacity of the network and taking into account the forecasted load due to EVs charging.

Comparative study on credit card fraud detection based on different support vector machines

Credit card fraud is the new financial fraud crime accompanied by the gradual development of the economy which causes billions of dollars of losses every year. Credit card fraud case not only seriously violated the cardholder benefits and financial institutions, but also undermined the credit management order. However, fraudsters keep exploring new crime strategies constantly which exacerbates the crime rate of fraud. Thus, a predictive model for credit card fraud detection is essential to minimize its losses. By distinguishing between fraud and non-fraud, machine learning is one of the most efficient solutions for detecting fraud. Support vector machines have proven to be a novel algorithm with excellent performance. Nevertheless, the performance of SVM depends largely on the correct choice of model parameters (C and g), which could cause that the false positive was very high if the kernel function type and parameter cannot be selected properly. In this paper, based on the real transaction data of the credit card business, firstly, it will find the optimal kernel function suitable for the data set. Secondly, this paper will propose the method of optimizing the support vector machine parameters by the cuckoo search algorithm, genetic algorithm and particle swarm optimization algorithm. Last but not least, the Linear kernel function was found to be the best kernel function with an accuracy rate of 91.56%. Furthermore, the Radial basis function is used to optimize the kernel function, which can improve the accuracy from 42.86% to the highest accuracy rate of 98.05%. Compared with CS-SVM and GA-SVM, PSO-SVM has the best overall performance.

The mathematically best way to spend your money: an algorithm for optimising the cashback profit

Cashback is one of the most popular ways for a bank to encourage its clients to use its financial services. By choosing a debit or a credit card wisely, one can greatly increase the corresponding cashback profit. However, if one is actively using two or more bank cards, it is not clear how exactly to divide one’s spendings across these cards in order to maximise the profit. In this work we provide an exact algorithm for this task, based on a known exact algorithm of linear programming. The proposed method is tested on real transactional data by comparing the historical cashback profit to some of the most profitable solutions provided by the proposed algorithm.

Deep Learning and Explainable Artificial Intelligence Techniques Applied for Detecting Money Laundering–A Critical Review

Money laundering has been a global issue for decades, which is one of the major threat for economy and society. Government, regulatory and financial institutions are combating it together in their respective capacity, however still billions of dollars in fines by authorities make the headlines in the news. High-speed internet services have enabled financial institutions to deliver better customer experience through multi-channel engagements, which has led to exponential growth in transactions and new avenues for laundering the money for fraudsters. Literature shows the usage of statistical methods, data mining and Machine Learning (ML) techniques for money laundering detection, but limited research on Deep Learning (DL) techniques, primarily due to lack of model interpretability and explainability of the decisions made. Several studies are conducted on application of ML for Anti-Money Laundering (AML), and Explainable Artificial Intelligence (XAI) techniques in general, but lacks the study on usage of DL techniques together with XAI. This paper aims to review the current state-of-the-art literature on DL together with XAI for identifying suspicious money laundering transactions and identify future research areas. Key findings of the review are, researchers have preferred variants of Convolutional Neural Networks, and AutoEncoder; graph deep learning together with natural language processing is emerging as an important technology for AML; XAI use is not seen in AML domain; 51% ML methods used in AML are non-interpretable, 58% studies used sample of old real data; key challenges for researchers are access to recent real transaction data and scarcity of labelled training data; and data being highly imbalanced. Future research directions are, application of XAI techniques to bring-out explainability, graph deep learning using natural language processing (NLP), unsupervised and reinforcement learning to handle lack of labelled data; and joint research programs between research community and industry to benefit from domain knowledge and controlled access to data.

Applying Comparable Sales Method to the Automated Estimation of Real Estate Prices

In this paper, we propose a novel procedure designed to apply comparable sales method to the automated price estimation of real estates, in particular, that of apartments. Apartments are the most popular residential housing type in Korea. The price of a single apartment is influenced by many factors, making it hard to estimate accurately. Moreover, as an apartment is purchased for living, with a sizable amount of money, it is mostly traded infrequently. Thus, its past transaction price may not be particularly helpful to the estimation after a certain period of time. For these reasons, the up-to-date price of an apartment is commonly estimated by certified appraisers, who typically rely on comparable sales method (CSM). CSM requires comparable properties to be identified and used as references in estimating the current price of the property in question. In this research, we develop a procedure to systematically apply this procedure to the automated estimation of apartment prices and assess its applicability using nine years’ real transaction data from the capital city and the most-populated province in South Korea and multiple scenarios designed to reflect the conditions of low and high fluctuations of housing prices. The results from extensive evaluations show that the proposed approach is superior to the traditional approach of relying on real estate professionals and also to the baseline machine learning approach.

No Place Like Home: Cross-National Data Analysis of the Efficacy of Social Distancing During the COVID-19 Pandemic.

BackgroundIn the absence of a cure in the time of a pandemic, social distancing measures seem to be the most effective intervention to slow the spread of disease. Various simulation-based studies have been conducted to investigate the effectiveness of these measures. While those studies unanimously confirm the mitigating effect of social distancing on disease spread, the reported effectiveness varies from 10% to more than 90% reduction in the number of infections. This level of uncertainty is mostly due to the complex dynamics of epidemics and their time-variant parameters. However, real transactional data can reduce uncertainty and provide a less noisy picture of the effectiveness of social distancing.ObjectiveThe aim of this paper was to integrate multiple transactional data sets (GPS mobility data from Google and Apple as well as disease statistics from the European Centre for Disease Prevention and Control) to study the role of social distancing policies in 26 countries and analyze the transmission rate of the coronavirus disease (COVID-19) pandemic over the course of 5 weeks.MethodsRelying on the susceptible-infected-recovered (SIR) model and official COVID-19 reports, we first calculated the weekly transmission rate (β) of COVID-19 in 26 countries for 5 consecutive weeks. Then, we integrated these data with the Google and Apple mobility data sets for the same time frame and used a machine learning approach to investigate the relationship between the mobility factors and β values.ResultsGradient boosted trees regression analysis showed that changes in mobility patterns resulting from social distancing policies explain approximately 47% of the variation in the disease transmission rates.ConclusionsConsistent with simulation-based studies, real cross-national transactional data confirms the effectiveness of social distancing interventions in slowing the spread of COVID-19. In addition to providing less noisy and more generalizable support for the idea of social distancing, we provide specific insights for public health policy makers regarding locations that should be given higher priority for enforcing social distancing measures.