Automated Induction of Rule-based Neural Networks from Databases

This paper describes our approach to the problem of automated knowledge acquisition from large databases of examples using an information-theoretic approach. Our previous research has resulted in practical algorithms (ITRULE) for the automatic induction of rules from large example databases. Utilizing these algorithms, the raw data can be transformed into a set of human readable IF THEN rules, thus giving insight into the knowledge hidden within the data. These rules can then be automatically loaded into an expert system shell. Alternatively, they can be used to build a new type of parallel inference system—a rule-based neural network. This process enables a prototype expert system to be automatically generated and up and running in a matter of minutes, compared with months using a manual knowledge-acquisition approach. The resulting expert system can then be used as a sophisticated search and analysis tool to query the original database capable of reasoning with uncertain and incomplete data.     

Effects of database choice on international accounting research

Data availability is one of the traditional obstacles confronting researchers carrying out international empirical studies in accounting. In recent years several databases have claimed to offer comprehensive coverage of accounting and financial data of firms worldwide. We analyse whether the choice of database has an effect on the results of empirical studies. We find that the results of a simple empirical adaptation of the Ohlson (1995 ) model for fourteen member states of the European Union change significantly depending on the database chosen (Datastream, Global Vantage, Company Analysis, Worldscope, Thomson Financial, Financials and BvD Osiris). These differences are mainly attributable to differences in the samples across databases. When we match observations across all databases the differences persist but are much less pronounced. Our main conclusion is that database choice matters, as it leads to different results when the same research design is used.     

Further evidence of the effects of normalization on end-user query errors: an experimental evaluation

The proliferation of Windows-based programs for PCs, together with the requirement of many business managers for faster responses to their information needs, has led many business end-users to create, maintain, and query their own databases. These individuals use the output of these queries as the basis for operational, tactical, and strategic decisions. To maximize the benefit of end-users' time, querying these databases must be as efficient as possible. Of even greater importance, the effectiveness of managers' decision-making is directly related to the quality of the information extracted. Because end-user querying is error prone, characterizing the sources of query errors and using that knowledge to improve the effectiveness of end-user query development can improve the quality of information available for decision-making.IS professionals typically design databases in third normal form. End-user databases designed with the help of IS professionals are also likely to be in third normal form. Although third normal form is better for data capture, prior research indicates third normal form is not necessarily the most appropriate normal form for querying. This paper reports the results of an in-depth experiment into the effects of normalization level on the efficiency and effectiveness of end-user querying. The results confirm earlier findings that end-users querying first normal form data structures were both more efficient and more effective than those querying third normal form data structures. This research extends prior research by examining the specific areas where first normal end-users outperform third normal form end-users. In particular, the experiment revealed that first normal form end-users made significantly fewer errors in relation to attribute selection, table selection, and row restriction.     

Which cryptocurrency data sources should scholars use?

Inspired by Alexander and Dakos (2020), we shed more light on the adequacy of data in the cryptocurrency literature by analysing the scaling properties and underlying processes of the main cryptocurrency databases (Coinmarketcap, Coingecko, BraveNewCoin and Cryptocompare) and exchange platforms (Coinbase, Bitstamp, Bittrex, Cexio and Exmo). Our results show that coin-ranking sites, such as Coinmarketcap, Coingecko and BraveNewCoin (i) include most of the cryptocurrency trading activity and (ii) are essentially characterised by the same underlying processes as the main exchange platforms (Coinbase and Bitstamp) and alternative coin-ranking sites (Cryptocompare), regardless of the possible issues arising from the aggregation of different exchanges to compute a unique cryptocurrency price. Therefore, we state that these databases are appropriate to conduct research. At any rate, we observe that all the databases analysed in this paper show the same underlying process for most liquid cryptocurrencies; consequently, scholars could use any of them for their studies, as long as they consider the different trading activity included by each database. This result is supported by an empirical analysis focused on weak-form market efficiency, since we report the same degree of efficiency regardless of the database and exchange platform. Nevertheless, we recognise the need for further research, given the gap in the literature and the black-box method used by coin-ranking sites to compute a unique cryptocurrency price.     

大数据时代宏观数据与微观数据的融合

在大数据时代,现有的数据质量参差不齐、缺少统一的代码和口径,使得宏观数据与微观数据之间的融合存在问题。以宏观数据库为基准,将微观数据库匹配到宏观数据库是解决数据融合难题的一个方法。建议设置宏观数据库与微观数据库,并鼓励学者上传数据、定期融合数据库、加强与科研机构和数据公司的合作,以此来完善我国数据库的建设工作。     

The Uses of Advanced Information Technology in Audit Planning

Expert systems, intelligent databases and other forms of advanced information technology are quickly becoming pervasive tools in accounting and auditing. All the major public accounting firms are either using such systems in their auditing practice or have them under development. This paper describes the use of several systems used for audit planning. These are divided into three functional areas: audit risk assessment, internal control evaluation and audit program development. The advantages and disadvantages of these systems as they relate to auditing are discussed.     

End-user access to multiple sources: incorporating knowledge discovery into knowledge management

The End-User Access to Multiple Sources—Eams system—integrates given information sources into a knowledge management system. It relates the world of documents with the database world using an ontology. The focus of developing the Eams system is on the acquisition and maintenance of knowledge. Hence, in both worlds, machine learning is applied. In the document world, a learning search engine adapts to user behaviour by analysing the click-through-data. This eases the personalization of selecting appropriate documents for users and does not require further maintenance. In the database world, knowledge discovery in databases (KDD) bridges the gap between the fine granularity of relational databases and the actual information needs of users. KDD extracts knowledge from data and, therefore, allows the knowledge management system to make good use of already existing company data—without further acquisition or maintenance. A graphical user interface provides users with a uniform access to document collections on the Internet (Intranet) as well as to relational databases. Since the ontology generates the items in the user interface, a change in the ontology automatically changes the user interface without further efforts. The Eams system has been applied to customer relationship management in the insurance domain. Questions to be answered by the system concern customer acquisition (e.g. direct marketing), customer up- and cross-selling (e.g. which products sell well together), and customer retention (here, which customers are likely to leave the insurance company or ask for a return of a capital life insurance). Documents about other insurance companies and demographic data published on the Internet contribute to the answers, as do the results of data analysis of the company's contracts. Copyright © 2003 John Wiley & Sons, Ltd.     

Integrating a CD-ROM database in the accounting curriculum

This paper presents a methodology for keeping students abreast of a new information technology by integrating the use of optical media for storage and distribution into the accounting curriculum. Two student projects are presented which build proficiency in locating and organizing data in a database and in using intellectual skills to analyze that data. The authors found that these projects made the course material more relevant and meaningful for both the students and the instructor.     

Addressing business needs: A creative module for teaching data macros in an accounting information systems course

Although database technology has advanced considerably, Accounting Information Systems (AIS) courses in business schools generally include only the basic aspects of databases, such as creating tables, queries, forms, reports and relationships between tables. One advanced topic that can be covered to enrich such courses is “triggered events.” Triggered events are useful in implementing accounting processes, such as data validation, control and exception reporting in databases. To date, teaching triggered events required enterprise-level database systems, thereby making it difficult to replicate in an introductory class. However, data macro features introduced in the 2010 version of Microsoft Access can now be used to teach such concepts in an easy-to-use interface. The proposed module in data macros helps students gain the ability to implement validity control and exception-reporting in databases. In addition, the module helps accounting students realize the importance of databases in meeting certain accounting needs. The effectiveness of the proposed module was assessed by an exercise and a survey.     

The twilight of the Baconian age and the future of humanity

Our times are the product of the unfolding of the Baconian program, articulated nearly 400 years ago by Sir Francis Bacon, which has been extraordinarily successful and has underpinned the worldwide expansion of Western civilization. However, in a paradoxical manner, the triumph of Bacon's program ended up undermining its own foundations. As we enter into the 21st century, we are witnessing the twilight of the Baconian age and we must embark in the search for a new program to mobilize human endeavors. This requires examining the logic and the underlying assumptions of the Baconian program, assessing the way in which they have been questioned by their own results and products, and exploring how to design and put in practice a new program to fully realize the human potential. This new program must build on the achievements of the Baconian age, but also acknowledge its limitations. This might, for example, call for putting emotions and feelings on a qualitative par with reason, and also for incorporating the contributions of non-Western cultural perspectives into the design of a new program for humanity.     

The effects of normalization on end-user query errors: An experimental evaluation

As a result of the shortage of professional programmers to extract timely information from databases, end-users are increasingly developing their own database queries. Because end-user querying is error-prone, characterizing the sources of query errors and using that knowledge to improve the effectiveness of end-user query development can improve the quality of information used for decision making. This paper reports the results of an experiment that investigated the effect of normalization level on query errors. The results show that query errors vary with the normalization level of the database structure and confirm previous findings about query errors increasing with task complexity. End-users querying a first normal form data structure make fewer errors than end-users querying an unnormalized data structure or a third normal form data structure. Furthermore, end-users querying a third normal form data structure make fewer errors than end-users querying an unnormalized data structure.     

Belief Networks for Expert System Development in Auditing

This study examines the use of a belief network based expert system for an auditing task—financial distress evaluation for banks. A belief network uses probability measures to store important dependencies across variables of interest in a problem domain, and makes inferences based on observed evidence using probability calculus. This paper discusses how belief network structures can be constructed, and used to assist auditor's in making appropriate recommendations regarding the financial health of a bank under audit. The ability of a belief network to make reliable predictions depends on how well the network structure reflects the underlying dependencies across variables in the problem domain (e.g. financial ratios and the financial health of a bank). The first part of this study illustrates how a computer program developed by the authors can be used to generate and evaluate different feasible belief network structures based on historical data. The program uses an information-theoretic measure to compare the alternative structures. The ability of the program to identify existing dependencies across variables is demonstrated by using it to reconstruct a known network structure from simulated data. Next, the program is used on a database of twelve important bank financial ratios over a three-year period. The predictive ratios identified by the program reflect important areas of a bank's health, such as loan quality, efficiency, profitability and capital adequacy. Finally, a belief revision mechanism is encoded for the belief network structure identified earlier, and is used to illustrate how it can assist auditors in making recommendations about financial health based on a bank's critical financial ratios. The probability estimates provided by the system are validated using data on banks not used in the network design stage, and are found to be reliable.     

MOMENTUM: DOES THE DATABASE MAKE A DIFFERENCE?

We examine discrepancies between the Center for Research in Security Prices (CRSP) and Trade and Quote (TAQ) databases by examining the returns of momentum strategies using each database. Momentum portfolios constructed from CRSP prices earn significant profits whereas similar portfolios using TAQ prices show losses. Adjusting TAQ prices with the TAQ dividends file or with the cumulative distribution factor provided by CRSP does not eliminate all differences. There are significant discrepancies in the way CRSP and TAQ record newly listed and delisted stocks. We document the residual (after all filters) price differences between the two databases and provide filters to adjust TAQ data for long sample periods and large sample sizes. Our filtering procedures allow for the possibility of examining intraday patterns in momentum profits.     

Some issues in blockchain for accounting and the supply chain,with an application of distributed databases to virtual organizations

This paper reviews some recent blockchain‐based applications for information capture, distribution and preservation. As part of that review, this paper examines two key concerns with current blockchain designs for accounting and supply chain transactions: data independence and multiple semantic models for the same information distribution problem. Blockchain applications typically integrate database, application and presentation tiers all in the same ledger. This results in a general inability to query information in the ledger and other concerns. Further, since most applications appear to be private blockchain applications, there is a concern of agents needing to accommodate multiple blockchains depending on who their trading partners are and what they request. Finally, this paper uses a distributed database to design a ‘blockchain‐like’ system for virtual organizations.     

The path to process data

Analysis of transformation processes is a potent tool for the understanding of materials utilization, energy consumption, environmental pollution and the productivity of capital and labour. This paper introduces process analysis with emphasis on the need for physical rather than monetary data. Some uncoordinated data gathering activities relevant to process analysis are examined. Two prototype process databases are described to illustrate ways of approaching process analysis and it is noted that a process database is necessary for the next generation of economic models. Although there is as yet little progress in assembling the organization to coordinate the building of such a database, some strategies for the future are proposed.     

Evaluating the SDC Mergers and Acquisitions Database

We compare 20 years of data from Thompson Financial SDC Platinum (SDC)'s Mergers and Acquisitions database with a hand‐collected database, providing evidence on the completeness and accuracy of SDC data across time. We find that our hand‐collected data is generally more accurate than SDC, but SDC's accuracy and coverage improves over time. Our investigation of discrepancies between the databases finds that SDC is more prone to errors on smaller, high book‐to‐market acquirers with weak announcement period market responses. Preliminary analyses suggest that this potential bias is not significant, but could affect inferences when examining smaller, high book‐to‐market firms.     

Science and policy in designing complex futures

Many complex systems scientists are motivated by making changes to socio-technical systems. Change also motivates policy makers. Thus, both aspire to design new systems and trajectories through time. Science has always been used to forecast the implications of policy. Traditional science predicts system states at precise points in time. Usually point predictions are impossible in the science of complex systems. Most scientists cannot perform in vivo experiments on complex socio-technical systems. They have neither the mandate nor the budget. Massive advances in ICT enable simulation as a powerful new in silico experimental methodology. However, complex systems scientists are unable to collect the huge databases necessary to plan and manage heterogeneous multilevel socio-technical systems—these data are collected by public and private agencies according to perceived policy needs. Most implemented policies are experiments, but the outcome of policy is rarely monitored from a scientific research perspective. To test in silico forecasts against in vivo observation requires data owned by policy makers, requiring completely new data collection protocols that have to be aligned with policy makers. Complex systems scientists must make an obvious added-value contribution to policy, making explicit their role as designers with their science fitting into the policy-driven process of designing, planning, managing and controlling real complex socio-technical systems. Remaining in the loop is essential for policy makers to maintain control over policy objectives as they co-evolve towards the delivered design solution and its implementation.     

An Empirical Investigation of Market Structure, Efficiency, and Performance in Property-Liability Insurance

This study examines the relationships among market structure and performance in property‐liability insurers over the period 1992–1998 using data at the company and group levels. Three specific hypotheses are tested: traditional structure‐conduct‐performance, relative market power, and efficient structure (ES). The results provide support for the ES hypothesis. The ES hypothesis posits that more efficient firms can charge lower prices than competitors, enabling them to capture larger market shares and economic rents, leading to increased concentration. Both revenue and cost efficiency are used in the analysis, and this is the first study to use revenue efficiency in this type of analysis. The results for the sample period as a whole and by year are consistent. The overall results suggest that cost‐efficient firms charge lower prices and earn higher profits, in conformance with the ES hypothesis. On the other hand, prices and profits are found to be higher for revenue‐efficient firms. Revenue X‐efficiency is derived from activities such as cross‐selling and may rely heavily on the use of detailed information from customer databases to identify potential customers. The implications of this research are that regulators should be more concerned with efficiency (both cost and revenue) rather than the market power that arises from the consolidation activity taking place in insurance.     

商业银行股权激励的特殊意义与现实问题分析

银行业普遍存在的管制,使银行面临着有限的控制权市场和接管威胁,从而大大降低了市场约束银行管理者的能力,因此对银行而言,通过股权激励机制协调股东和管理者的利益、降低代理成本显得尤为重要.我国商业银行实施股权激励不应盲目照搬西方国家的范例,应慎重选择股权激励的实施范围,目前最好只局限在高级管理人员,通过对股权激励机制的合理设计来充分发挥其长期激励效应,切不可盲目扩大激励范围,将股权激励演变成一种新的福利措施.应尽快建立起以经济资本为核心的业绩考核体系,在经济资本的约束下实现业务发展模式和盈利模式的转变.股权激励机制的有效实施离不开完善的银行治理结构.