Intelligent Systems in Accounting, Finance and Management最新文献

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.

In 2022, global startup investments exceeded US$445 billion, sourced from entities like venture capital (VC) funds, angel investors, and equity crowdfunding. Despite their role in driving innovation, startup investments often fall short of S&P 500 returns. Surprisingly, the potential of artificial intelligence (AI) remains untapped by investors, despite AI's growing sway in financial decision-making. Our empirical analysis predicts the success of 10,000 Israeli startups, utilizing diverse machine learning models. Unlike prior research, we employ the MetaCost algorithm to convert models into cost-sensitive variants, minimizing total cost instead of total error. This innovative approach enables varied costs linked to different prediction errors. Our results underscore that these cost-sensitive machine learning models significantly reduce risk for VC funds and startup investors compared to traditional ones. Furthermore, these models provide investors with a distinct capability to tailor their risk profiles, aligning predictions with their risk appetite. However, while cost-sensitive machine learning reduces risk, it may limit potential gains by predicting fewer successful startups. To address this, we propose methods to enhance successful startup identification, including aggregating outcomes from multiple MetaCost models, particularly advantageous for smaller deal flows. Our research advances AI's role in startup investing, presenting a pivotal tool for investors navigating this domain.

There has been substantial discussion aimed at investigating the extent to which academic researchers can or should “use” large language models, such as ChatGPT and Bard, in their research papers. However, there seems to have been limited attention given to the extent to which students can use these tools for the development of theses, proposals and dissertations. This paper pushes the arguments from focusing on academic researchers, journal papers, and technical meetings to considering those theses and dissertations, raising several questions and concerns. Ultimately, university policies need to address these issues, but if publisher and editor responses and alternative business uses are a signal of that direction, consensus may be difficult to achieve.

This paper provides a first comprehensive definition of cryptoassets for accounting purposes in the types of payment tokens, electronic money (e-money) tokens, utility tokens and security tokens. The delivery of definitions for accounting purposes addresses some of the concerns raised by the European Financial Reporting Advisory Group (EFRAG) discussion paper and helps accounting regulators adapt current International Financial Reporting Standards (IFRS) standards to blockchain-based tokens' taxonomy and nature. The paper helps policymakers reconcile Markets in Cryptoassets Regulation Proposal's (MiCA) definitions and classification of cryptoassets with the EFRAG's specific needs for clarification and/or amendment in the IFRS standards and contributes to providing an accounting guide for practitioners in their financial disclosure.

We deal with a multivariate conditional value at risk. Compared with the usual notion for the single random variable, a multivariate value at risk is concerned with several variables, and thus, the relation between each risk factor should be considered. We here introduce a new definition of copula-based conditional value at risk, which is real valued and ready to be computed. Copulas are known to provide a flexible method for handling a possible nonlinear structure; therefore, copulas may be naturally involved in the theory of value at risk. We derive a formula of our copula-based conditional value at risk in the case of Archimedean copulas, whose effectiveness is shown by examples. Numerical studies are also carried out with real data, which can be verified with analytical results.

Comparing companies can be useful for various purposes. Despite the widespread use of industry classification systems as a peer selection standard, these have been criticized for various reasons. Financial statements, however, offer a promising alternative to such classification systems. They are standardized, widely available, and offer deep insights into the nature of the company. In this paper, we present a graph distance metric for financial statements using the earth mover's distance. When using the distance metric on real-world tasks such as peer identification and industry classification, it shows promising results in terms of accuracy and computational efficiency.