Intelligent Systems in Accounting, Finance and Management最新文献

Metaverse is an emerging digital space that uses innovative technologies to allow users to facilitate building relationships virtually and to create new interaction opportunities. Even, the financial sector has been disrupted by the metaverse involving digital assets, cryptocurrencies, blockchain technology, and decentralized finance. The objective of this paper is to focus on novel intelligent systems technologies with the potential for application in the financial area to have a better knowledge of the current research topics, challenges, and future directions. A systematic literature review was conducted analyzing papers on technological innovation of the metaverse in financial sector. Following the PRISMA methodology, we have selected 29 primary studies from five scientific databases to be included in the review. The results show that 11 types of innovative metaverse technologies are applied in the financial sector, developing financial innovations, among which the most discussed is cryptocurrency. Among the opportunities that the use of the metaverse brings to the financial sector, the reduction of transaction costs is the most discussed. Finally, five open challenges in the use of metaverse technologies in the financial sector have been identified, relating to the use of data, the application of technologies, social integration, financial innovation, and regulatory compliance. Based on this study, recommendations on future research directions are provided to the scientific community.

Liquidity planning and forecasting are essential activities in corporate financial planning team. Traditionally, empirical models and techniques based on in-house expertise have been used to navigate the numerous challenges of this forecasting activity. These challenges become more complex when the forecasting activities are extended to subsidiaries of a large firm. This paper presents a structured approach that utilizes 240 covariates to predict net liquidity, customer receipts, and payments to suppliers to improve the accuracy and efficiency of liquidity forecasting in subsidiaries and at the corporate level. The approach is empirically validated on a large corporation headquartered in Germany, with average annual revenue from 6 to 7 billion Euro spanning 80 countries. The proposed approach demonstrated superior performance over existing methods in six out of nine forecasts using the data from 2014 to 2018. These findings suggest that a firm's classical approach to liquidity forecasting can be effectively challenged and outperformed by the algorithmic approach.

Although the financial audit controls in companies have advanced over the years, the number of corporate fraud instances is growing, thus raising the need for investigating the factors that can be used as early warning signals and developing effective systems for identifying financial fraud. In this paper, financial statements from 133 Greek companies listed in the Athens Stock Exchange over the period 2014 to 2019 are investigated, based on the fraud diamond theory. Financial data and corporate governance variables are used as inputs to data mining techniques to develop models that can identify patterns of irregularities in a company's financial reports. To this end, popular machine learning classification algorithms are employed in a novel multi-label classification setting that not only identifies fraudulent cases but also considers the nature of the auditors' comments. The results indicate that the proposed multi-label approach provides enhanced results compared to binary classification algorithms, avoiding inconsistent outputs with respect to the existence of different forms of manipulation of financial statements.

We investigate applying and extending an exhaust data framework, using an empirical analysis to explore and compare different predictive analytic capabilities of both internal and external exhaust data for estimating sales. We use internal exhaust data that explores the relationship between app usage and web traffic data and estimation of sales and find the ability to predict sales at least 4 days ahead. We also develop predictive models of sales, using external data of Google searches, extending the previous research to include additional macroeconomic Google variables and Wikipedia pageviews, finding that we can predict at least 4 months ahead, suggesting a portfolio of exhaust data be used. We introduce the roles of internal and external exhaust data, direct and indirect exhaust data and transformed exhaust data, into an exhaust data framework. We examine what appear to be different levels of information fineness and predictability from those exhaust data sources. We also note the importance of the types of devices (e.g., mobile) and the types of commerce (e.g., mobile commerce) in creating and finding different types of exhaust. Finally, we apply an existing exhaust data framework to develop macroeconomic data exhaust variables, as the means of capturing inflation and unemployment information, using Google searches.

Modern financial exchanges use an electronic limit order book (LOB) to store bid and ask orders for a specific financial asset. As the most fine-grained information depicting the demand and supply of an asset, LOB data are essential in understanding market dynamics. Therefore, realistic LOB simulations offer a valuable methodology for explaining the empirical properties of markets. Mainstream simulation models include agent-based models (ABMs) and stochastic models (SMs). However, ABMs tend not to be grounded on real historical data, whereas SMs tend not to enable dynamic agent-interaction. More recently, deep generative approaches have been successfully implemented to tackle these issues, whereas its black-box essence hampers the explainability and transparency of the system. To overcome these limitations, we propose a novel hybrid neural stochastic agent-based model (NS-ABM) for LOB simulation that incorporates a neural stochastic trader in agent-based simulation, characterised by (1) representing the aggregation of market events' logic by a neural stochastic background trader that is pre-trained on historical LOB data through a neural point process model; (2) learning the complex distribution for order-related attributes conditioned on various market indicators through a non-parametric diffusion probabilistic model; and (3) embedding the background trader in a multi-agent simulation platform to enable interaction with other strategic trading agents. We instantiate this hybrid NS-ABM model using the ABIDES platform. We first run the background trader in isolation and show that the simulated LOB can recreate a comprehensive list of stylised facts that demonstrate realistic market behaviour. We then introduce a population of ‘trend’ and ‘value’ trading agents, which interact with the background trader. We show that the stylised facts remain and we demonstrate order flow impact and financial herding behaviours that are in accordance with empirical observations of real markets.

Directional change (DC) is an alternative to time series in recording transactions: it only records the transactions at which price changes to the opposite direction of the current trend by a threshold specified by the observer. DC can only be confirmed in hindsight: one does not know that direction has changed until it is confirmed by a later transaction. The transaction in which the price confirms a DC is called a DC confirmation point. DC nowcasting is an attempt to recognize DC before the DC confirmation point. Accurate DC nowcasting will benefit trading. In this paper, we propose a method for DC nowcasting. This method is entirely data driven: it is based on the historical distribution of DC-related indicators. Empirical results suggest that DC nowcasting is possible, even under a naïve rule. This opens the door to a promising research direction on an important topic.

Less than 2 years ago, many small entrepreneurs in the commodities trading business faced price volatility, which had not been the case for the last few decades. Generally, the income section of the profit and loss statement was not the main problem, especially in building material commodities trading, due to the recent growth in real estate demand. Logistic disorders, raw material shortages, inflation, and interest rate growth caused difficulties in supply management and warehouse balancing, which were reflected in a particular significant expense called the cost of goods sold. The real problem of its forecasting was identified, and data from accounting books likely contain information about previous warehouse dynamics. This paper presents how accounting data are prepared and shaped into time series suitable for machine learning algorithms, the relevant literature that helped in algorithm selection, and the development and description of the forecasting model, as well as its benchmarking with traditional forecasting models. Visualization and mean squared error loss measured on unseen data show that the model has proven more successful than expected. Based on data from four journal accounts spanning over 14 years, the model predicts the debit and credit sides of the wholesale warehouse for 150 working days.

Advances in Deep Learning have drastically improved the abilities of Natural Language Processing (NLP) research, creating new state-of-the-art benchmarks. Two research streams at the forefront of NLP analysis are transformer architecture and multimodal analysis. This paper critically evaluates the extant literature applying sentiment analysis techniques to the financial domain. We classify the financial sentiment analysis literature according to the most used techniques in the area, with a focus on methods used to detect sentiment within corporate earnings conference calls, because of their dual modality (text-audio) nature. We find that the financial literature follows a similar path to NLP sentiment literature, in that more advanced techniques to define sentiment are being used as the field progresses. However, techniques used to determine financial sentiment currently fall behind state-of-the-art techniques used within NLP. Two future directions stem from this paper. Firstly, we propose that the adoption of transformer architecture to create robust representations of textual data could enhance sentiment analysis in academic finance. Secondly, the adoption of multimodal classifiers in finance represents a new, currently underexplored area of study that offers opportunities for finance research.