Search

This article presents a comprehensive methodology for processing financial datasets of Apple Inc., encompassing quarterly income and daily stock prices, spanning from March 31, 2009, to December 31, 2023. Leveraging 60 observations for quarterly income and 3774 observations for daily stock prices, sourced from Macrotrends and Yahoo Finance respectively, the study outlines five distinct datasets crafted through varied preprocessing techniques. Through detailed explanations of aggregation, interpolation (linear, polynomial, and cubic spline) and lagged variables methods, the study elucidates the steps taken to transform raw data into analytically rich datasets. Subsequently, the article delves into regression analysis, aiming to decipher which of the five data processing methods best suits capital market analysis, by employing both linear and polynomial regression models on each preprocessed dataset and evaluating their performance using a range of metrics, including cross-validation score, MSE, MAE, RMSE, R-squared, and Adjusted R-squared. The research findings reveal that linear interpolation with polynomial regression emerges as the top-performing method, boasting the lowest validation MSE and MAE values, alongside the highest R-squared and Adjusted R-squared values.

Accounting informatization is a crucial component of enterprise informatization, significantly impacting operational efficiency in accounting and finance. Advances in information technology have introduced automation techniques that accelerate the processing of accounting information cost-effectively. Integrating artificial intelligence, cloud computing, and edge computing is pivotal in streamlining and optimizing these processes. Traditionally, accounting informatization relied on system servers and local storage for data processing. However, the era of big data necessitates a shift to cloud computing frameworks for efficient data storage and processing. Despite the advantages of cloud storage, concerns arise regarding data security and the substantial data transactions between the cloud and source devices. To address these challenges, this research proposes a novel algorithm, Heterogeneous Distributed Deep Learning with Data Offloading (DDLO) algorithm. DDLO leverages the synergy between edge devices and cloud computing to enhance data processes. Edge computing enables rapid processing of large volumes of data at or near the data collection sites, optimizing day-to-day operations for enterprises. Furthermore, machine learning algorithms at edge devices enhance data processing efficiency, augmenting the computing environment's overall performance. The proposed DDLO algorithm fosters a hybrid machine learning approach for computing joint tasks and multi-tasking in accounting informatization. It enables dynamic resource allocation, allowing selected data or model updates to be offloaded to the cloud for complex tasks. The algorithm's performance is rigorously evaluated using key metrics, including computing time, offloading time, accuracy, and cost levels. By capitalizing on the strengths of edge computing, cloud computing, and artificial intelligence, the DDLO algorithm effectively addresses accounting informatization challenges. It empowers enterprises to process vast amounts of accounting data efficiently and securely while improving overall operational efficiency. Regarding time, using terasort in tasks offloading using DDLO consumes less milliseconds 0t 33 ms which is lesser than other techniques. [ABSTRACT FROM AUTHOR]

While Machine Learning significantly boosts the performance of predictive models, its efficacy varies across different data dimensions. It is essential to cluster time series data of similar characteristics, particularly in the financial sector. However, clustering financial time series data poses considerable challenges due to the market’s inherent complexity and multidimensionality. To address these issues, our study introduces a novel clustering framework that leverages autoencoders for a compressed yet informative representation of financial time series. We rigorously evaluate our approach through multiple dimensionality reduction and clustering algorithms, applying it to key financial indices, including IBEX-35, CAC-40, DAX-30, S&P 500, and FTSE 100. Our findings consistently demonstrate that incorporating autoencoders significantly enhances the granularity and quality of clustering, effectively isolating distinct categories of financial time series. Our findings carry significant ramifications for the financial industry. By refining clustering methodologies, we set the stage for increasingly accurate financial predictive models, offering valuable insights for optimizing investment strategies and enhancing risk management.

Abstract The Stock Market is one of the most active research areas, and predicting its nature is an epic necessity nowadays. Predicting the Stock Market is quite challenging, and it requires intensive study of the pattern of data. Specific statistical models and artificially intelligent algorithms are needed to meet this challenge and arrive at an appropriate solution. Various machine learning and deep learning algorithms can make a firm prediction with minimised error possibilities. The Artificial Neural Network (ANN) or Deep Feed‐forward Neural Network and the Convolutional Neural Network (CNN) are the two network models that have been used extensively to predict the stock market prices. The models have been used to predict upcoming days' data values from the last few days' data values. This process keeps on repeating recursively as long as the dataset is valid. An endeavour has been taken to optimise this prediction using deep learning, and it has given substantial results. The ANN model achieved an accuracy of 97.66%, whereas the CNN model achieved an accuracy of 98.92%. The CNN model used 2‐D histograms generated out of the quantised dataset within a particular time frame, and prediction is made on that data. This approach has not been implemented earlier for the analysis of such datasets. As a case study, the model has been tested on the recent COVID‐19 pandemic, which caused a sudden downfall of the stock market. The results obtained from this study was decent enough as it produced an accuracy of 91%.

Abstract Financial fraud arises from the exaggeration of business interests, and an accurate detection or prediction is a useful tool for both corporate management and capital market systems. A collection of computer technologies has been made on this problem so far, and one of the most important solutions is unsupervised learning algorithms. Among them, most approaches work by analysing the internal relations in financial data and finding a new description of non‐fraud firms. However, current studies focus a lot on the geometry attribute of financial data, while overlooking the obvious behaviour patterns and peer effects among firms. This has limited the accuracy of representation and furthermore the detection performance. In this work, a very general class of functions is allowed to represent firms, constraining them by peer effects between firms and presenting an error‐distribution‐based financial fraud firm detection approach. Experimental results have shown great performance of the proposed approach.

The Stock Market is one of the most active research areas, and predicting its nature is an epic necessity nowadays. Predicting the Stock Market is quite challenging, and it requires intensive study of the pattern of data. Specific statistical models and artificially intelligent algorithms are needed to meet this challenge and arrive at an appropriate solution. Various machine learning and deep learning algorithms can make a firm prediction with minimised error possibilities. The Artificial Neural Network (ANN) or Deep Feed‐forward Neural Network and the Convolutional Neural Network (CNN) are the two network models that have been used extensively to predict the stock market prices. The models have been used to predict upcoming days' data values from the last few days' data values. This process keeps on repeating recursively as long as the dataset is valid. An endeavour has been taken to optimise this prediction using deep learning, and it has given substantial results. The ANN model achieved an accuracy of 97.66%, whereas the CNN model achieved an accuracy of 98.92%. The CNN model used 2‐D histograms generated out of the quantised dataset within a particular time frame, and prediction is made on that data. This approach has not been implemented earlier for the analysis of such datasets. As a case study, the model has been tested on the recent COVID‐19 pandemic, which caused a sudden downfall of the stock market. The results obtained from this study was decent enough as it produced an accuracy of 91%. [ABSTRACT FROM AUTHOR]

Abstract Predicting the correct values of stock prices in fast fluctuating high‐frequency financial data is always a challenging task. A deep learning‐based model for live predictions of stock values is aimed to be developed here. The authors' have proposed two models for different applications. The first one is based on Fast Recurrent Neural Networks (Fast RNNs). This model is used for stock price predictions for the first time in this work. The second model is a hybrid deep learning model developed by utilising the best features of FastRNNs, Convolutional Neural Networks, and Bi‐Directional Long Short Term Memory models to predict abrupt changes in the stock prices of a company. The 1‐min time interval stock data of four companies for a period of one and three days is considered. Along with the lower Root Mean Squared Error (RMSE), the proposed models have low computational complexity as well, so that they can also be used for live predictions. The models' performance is measured by the RMSE along with computation time. The model outperforms Auto Regressive Integrated Moving Average, FBProphet, LSTM, and other proposed hybrid models on both RMSE and computation time for live predictions of stock values.

Financial fraud arises from the exaggeration of business interests, and an accurate detection or prediction is a useful tool for both corporate management and capital market systems. A collection of computer technologies has been made on this problem so far, and one of the most important solutions is unsupervised learning algorithms. Among them, most approaches work by analysing the internal relations in financial data and finding a new description of non‐fraud firms. However, current studies focus a lot on the geometry attribute of financial data, while overlooking the obvious behaviour patterns and peer effects among firms. This has limited the accuracy of representation and furthermore the detection performance. In this work, a very general class of functions is allowed to represent firms, constraining them by peer effects between firms and presenting an error‐distribution‐based financial fraud firm detection approach. Experimental results have shown great performance of the proposed approach. [ABSTRACT FROM AUTHOR]

Abstract The Hipster Effect is a group of evolutionary ‘‘Diffusive Learning’’ processes of networks of individuals and groups (and their communication devices) that form Cyber‐Physical Systems; and the Hipster Effect theory has potential applications in many fields of research. This study addresses decision‐making parameters in machine‐learning algorithms, and more specifically, critiques the explanations for the Hipster Effect, and discusses the implications for portfolio management and corporate bankruptcy prediction (two areas where AI has been used extensively). The methodological approach in this study is entirely theoretical analysis. The main findings are as follows: (i) the Hipster Effect theory and associated mathematical models are flawed; (ii) some decision‐making and learning models in machine‐learning algorithms are flawed; (iii) but regardless of whether or not the Hipster Effect theory is correct, it can be used to develop portfolio management models, some of which are summarised herein; (iv) the [1] corporate bankruptcy prediction model can also be used for portfolio‐selection (stocks and bonds).

Forecasting of financial time series is an intensive working area for researchers and practitioners. In this study, we analyze 59 articles and discuss the progress in financial time series analysis using support vector machines. Our main conclusions are: (a) forecasting is doing in a daily basis and studies in other time scales are scarce; (b) most of works are devoted to improve the parameter estimation process or preprocessing the time series; (c) most of the work is concerned to forecast market financial indexes; (d) in general, is commonly accepted than support vector machines are more accurate than classical time series forecasting methods and other types of neural networks; collected evidence in this work, shows that modified versions of support vector machines are more accurate that the traditional version; however, it is not possible to conclude about what improvements are better because used experimental cases in each study are different and it is no possible to draw general conclusions. [ABSTRACT FROM PUBLISHER]

We implement a master-slave parallel genetic algorithm with a bespoke log-likelihood fitness function to identify emergent clusters within price evolutions. We use graphics processing units (GPUs) to implement a parallel genetic algorithm and visualise the results using disjoint minimal spanning trees. We demonstrate that our GPU parallel genetic algorithm, implemented on a commercially available general purpose GPU, is able to recover stock clusters in sub-second speed, based on a subset of stocks in the South African market. This approach represents a pragmatic choice for low-cost, scalable parallel computing and is significantly faster than a prototype serial implementation in an optimised C-based fourth-generation programming language, although the results are not directly comparable because of compiler differences. Combined with fast online intraday correlation matrix estimation from high frequency data for cluster identification, the proposed implementation offers cost-effective, near-real-time risk assessment for financial practitioners.

Abstract: In this paper we present an FPGA implementation of a Monte-Carlo method for pricing Asian options using Impulse C and floating-point arithmetic. In an Altera Stratix-V FPGA, a 149x speedup factor was obtained against an OpenMP-based solution in a 4-core Intel Core i7 processor. This speedup is comparable to that reported in the literature using a classic HDL-based methodology, but the development time is significantly reduced. Additionally, the use of a HLL-based methodology allowed us to implement a high-quality Gaussian random number generator, which produces more precise results than those obtained with the simple generators usually present in HDL-based designs. [Copyright &y& Elsevier]

After years of hype, the use of mobile devices to buy goods and services may be ready for prime time, assuming developers can overcome some obstacles. [ABSTRACT FROM AUTHOR]

The world of finance is an exciting and challenging environment. Recent years have seen an explosion in the application of computational intelligence methodologies in finance. In this article we provide an overview of some of these applications concentrating on those employing an evolutionary computation approach. [ABSTRACT FROM PUBLISHER]

The global economy is in worse shape. The credit crisis has spurred historic declines in property values, financial instruments that until recently traded freely have become illiquid, arresting the flow of assets; commodity prices gyrate wildly. We measure economic health by reading a wide range of indicators, from the growth rate of countries' GNPs to the volatility of multiple financial markets. And then there are the balance sheets of the world's banks. Finance is the cornerstone for the smooth functioning of the global economy. Unaware of the impact of their decisions, investors and traders large and small determine the relative prices of currencies, the level of interest rates, and stock prices. To develop predictive services that are reliable and add genuine value, we have to model not only the behavior of traders and their interactions but how trading activity relates to society in general. To create such robust models we need an event-based programming language. Olsen Routes software has been designed to enable the modeling of complex networks of interaction. [ABSTRACT FROM PUBLISHER]

The most common examples of financial variables are traded assets such as stocks and bonds. Derivatives have greatly facilitated financial management, the most useful technique of which is hedging, a trading activity that aims to reduce exposure to risk. The weather derivatives market has similarities to both the potential evolution and pricing of information security derivatives. Financial markets and investors often consider a cyberattack on a corporation's information systems to be a serious management problem and thus a threat to overall corporate value. This perception intensifies if the attack also involves direct theft of personal client data. In such a scenario, companies not only lose customer trust, but they're also subject to costly legal actions, both from customers and government regulators. Information security derivatives can provide a solution. The first is related to investors and brand name: a company that uses appropriate financial instruments to hedge its risks is likely to have investor trust and confidence. Prediction markets could assist in determining security attacks and their related risks as well as preferences. [ABSTRACT FROM PUBLISHER]

Fuzzy time-series models have been widely applied due to their ability to handle nonlinear data directly and because no rigid assumptions for the data are needed. In addition, many such models have been shown to provide better forecasting results than their conventional counterparts. However, since most of these models require complicated matrix computations, this paper proposes the adoption of a multivariate heuristic function that can be integrated with univariate fuzzy time-series models into multivariate models. Such a multivariate heuristic function can easily be extended and integrated with various univariate models. Furthermore, the integrated model can handle multiple variables to improve forecasting results and, at the same time, avoid complicated computations due to the inclusion of multiple variables. [ABSTRACT FROM AUTHOR]

An IT2 SFLS using hybrid RLS-BP training method was tested and compared for forecasting the daily exchange rate between Mexican Peso and U.S. Dollar (MXNUSD). The results showed that the IT2 SFLS forecaster using RLS-BP hybrid learning provided the best performance, and the base line T1 SFLS provided the worst performance. We conclude, therefore, that it is possible to directly use the daily data of exchange rate to train an IT2 SFLS, in order to predict MXNUSD exchange rate one day in advance. It was observed that IT2 SFLS forecasters efficiently managed the uncertainties presented in the raw historical data [ABSTRACT FROM PUBLISHER]

Purpose -- This paper is intended to test the robustness of the fitness of nested GARCH models. Design/methodology/approach -- Both Monte-Carlo simulation data and real-world data are used in the paper. Likelihood-family tests are used to test in-sample fitness, while mean-squared prediction error is employed for out-sample prediction tests. Findings -- The paper finds that, generally, the parsimonious principle is found to work well for both criteria. However, it is found that conflict exists between the two criteria: in-sample likelihood-family tests pay more attention to conditional distributions or are more sensitive to fat tail effects; while the out-sample criteria focus more on the accuracy of parameter estimation. Originality/value -- The paper shows that complexity does not necessarily mean good fitness; sometimes, the simpler model can fit better, especially for real-world data. [ABSTRACT FROM AUTHOR]

A fuzzy time series data representation method based on the Japanese candlestick theory is proposed and used in assisting financial prediction. The Japanese candlestick theory is an empirical model of investment decision. The theory assumes that the candlestick patterns reflect the psychology of the market, and the investors can make their investment decision based on the identified candlestick patterns. We model the imprecise and vague candlestick patterns with fuzzy linguistic variables and transfer the financial time series data to fuzzy candlestick patterns for pattern recognition. A fuzzy candlestick pattern can bridge the gap between the investors and the system designer because it is visual, computable, and modifiable. The investors are not only able to understand the prediction process, but also to improve the efficiency of prediction results. The proposed approach is applied to financial time series forecasting problem for demonstration. By the prototype system which has been established, the investment expertise can be stored in the knowledge base, and the fuzzy candlestick pattern can also be identified automatically from a large amount of the financial trading data. [ABSTRACT FROM AUTHOR]

The objective of this study is to explore ways of determining the useful lengths of intervals in fuzzy time series. It is suggested that ratios, instead of equal lengths of intervals, can more properly represent the intervals among observations. Ratio-based lengths of intervals are, therefore, proposed to improve fuzzy time series forecasting. Algebraic growth data, such as enrollments and the stock index, and exponential growth data, such as inventory demand, are chosen as the forecasting targets, before forecasting based on the various lengths of intervals is performed. Furthermore, sensitivity analyses are also carried out for various percentiles. The ratio-based lengths of intervals are found to outperform the effective lengths of intervals, as well as the arbitrary ones in regard to the different statistical measures. The empirical analysis suggests that the ratio-based lengths of intervals can also be used to improve fuzzy time series forecasting. [ABSTRACT FROM AUTHOR]

The authors analyze 40 online privacy policy documents from nine financial institutions to examine their clarity and readability. Their findings show that compliance with the existing legislation and standards is, at best, questionable. [ABSTRACT FROM PUBLISHER]