Search results

Social media provide an impressive amount of data about users and their interactions, thereby offering computer and social scientists, economists, and statisticians – among others – new opportunities for research. Arguably, one of the most interesting lines of work is that of predicting future events and developments from social media data. However, current work is fragmented and lacks of widely accepted evaluation approaches. Moreover, since the first techniques emerged rather recently, little is known about their overall potential, limitations and general applicability to different domains. Therefore, better understanding the predictive power and limitations of social media is of utmost importance.

Different types of forecasting models and their adaptation to the special circumstances of social media are analyzed and the most representative research conducted up to date is surveyed. Presentations of current research on techniques, methods, and empirical studies aimed at the prediction of future or current events from social media data are provided.

A taxonomy of prediction models is introduced, along with their relative advantages and the particular scenarios where they have been applied to. The main areas of prediction that have attracted research so far are described, and the main contributions made by the papers in this special issue are summarized. Finally, it is argued that statistical models seem to be the most fruitful approach to apply to make predictions from social media data.

This special issue raises important questions to be addressed in the field of social media-based prediction and forecasting, fills some gaps in current research, and outlines future lines of work.

This paper aims to explore the effectiveness of an artificial neural network (ANN) in predicting fraudulent financial reporting in small market capitalization companies in Malaysia.

Based on the concepts of ANN, a mathematical model was developed to compare non-fraud and fraud companies selected from among small market capitalization companies in Malaysia; the fraud companies had already been charged by the Securities Commission for falsification of financial statements. Ten financial ratios are used as fraud risk indicators to predict fraudulent financial reporting using ANN.

The findings indicate that the proposed ANN methodology outperforms other statistical techniques widely used for predicting fraudulent financial reporting.

The study is one of few to adopt the ANN approach for the prediction of financial reporting fraud.

The rapid development of e-commerce has brought not only great convenience to people but a great challenge to online stores. Phenomenon such as out of stock and slow sales has been common in recent years. These issues can be managed only when the occurrence of the sales volume is predicted in advance, and sufficient warnings can be executed in time. Thus, keeping in mind the importance of the sales prediction system, the purpose of this paper is to propose an effective sales prediction model and make digital marketing strategies with the machine learning model.

Based on the consumer purchasing behavior decision theory, we discuss the factors affecting product sales, including external factors, consumer perception, consumer potential purchase behavior and consumer traffic. Then we propose a sales prediction model, M-GNA-XGBOOST, using the time-series prediction that ensures the effective prediction of sales about each product in a short time on online stores based on the sales data in the previous term or month or year. The proposed M-GNA-XGBOOST model serves as an adaptive prediction model, for which the instant factors and the sales data of the previous period are the input, and the optimal computation is based on the proposed methodology. The adaptive prediction using the proposed model is developed based on the LSTM (Long Short-Term Memory), GAN (Generative Adversarial Networks) and XGBOOST (eXtreme Gradient Boosting). The model inherits the advantages among the algorithms with better accuracy and forecasts the sales of each product in the store with instant data characteristics for the first time.

The analysis using Jingdong dataset proves the effectiveness of the proposed prediction method. The effectiveness of the proposed method is enhanced and the accuracy that instant data as input is found to be better compared with the model that lagged data as input. The root means squared error and mean absolute error of the proposed model are found to be around 11.9 and 8.23. According to the sales prediction of each product, the resource can be arranged in advance, and the marketing strategy of product positioning, product display optimization, inventory management and product promotion is designed for online stores.

The paper proposes and implements a new model, M-GNA-XGBOOST, to predict sales of each product for online stores. Our work provides reference and enlightenment for the establishment of accurate sales-based digital marketing strategies for online stores.

The purpose of this paper is to subject the short lived “Policy Analysis Market” (PAM) – “a Pentagon betting market on terror attacks” – and media and academic reactions to it, to some critical analysis.

The paper engages sustained invocation of the relationship between simulation and representation, for the story of the Policy Analysis Market (PAM) and its demise is replete with the tension between the two. It interrogates a range of accounts of the (un)timely demise of PAM, from the fearful senators and the moralistic media who subsumed and buttressed their position to the market evangelists for whom the failure of this particular market was merely proof of the veracity of markets elsewhere.

It is found that, inter alia, PAM was not really market‐like enough and, indeed, that it duplicated in impoverished form already existing markets that pertain to its objects of interest; that it was too much a market, given that its “goods” are seemingly inappropriate for market trade; and that it exposed too much of the truth of the actual operation of existing markets via the difficulties it confronted with regard to the possibility of insider dealing.

By contextualising PAM within the so‐called war on terror of which it was part, we see in the tension between representation and simulation, tension between a singular and a manifold reality; a set of tensions which make clear the extent of the gap that must exist between cause and effect, truth and prediction. The paper concludes by joining the celebration of PAM's demise whilst yearning for a similar fate to befall the other monologues that brought it to silence.

The purpose of this paper is to contribute to the growing body of research in prediction markets by using trading data as a means of characterizing trader behavior in these markets. Traders are placed in homogenous groups based on common Trading habits using clustering algorithms. Several behavioral themes are used to guide the analyses.

Several market experiments were run to collect trading data, which was then exported into a data warehouse. A secondary data analysis is performed on variables derived from the original trade data. In particular, k‐means clustering is used to form groups of traders that share common characteristics.

Participants can be classified into homogenous groups based on their trading behavior. Groups tend to differ based on the overall level of participation, how much of their trading activity is spent buying or selling, and how early they enter the market.

More research should be done using a variety of variables to further determine the impact of various types of trader behavior on prediction markets.

Using insights gained from work like this, the design of prediction markets can be fine tuned to encourage behavior that contributes to trader participation and the overall accuracy of the market predictions.

Little research has been done to evaluate the impact of trader behavior on the accuracy of prediction markets. The authors used a new prediction market implementation to collect detailed trading data. This data was then used to derive higher level trading attributes that can he used to characterize traders. The k‐means clustering algorithm was shown to be an effective means of defining groups of traders who share common market behaviors.

The purpose of this paper is to propose an approach for predicting the movements of Chinese medicinal material price indexes using news based on text mining.

A research framework and three major methods, namely, domain dictionary construction, market convergence time calculation and dimensionality reduction integrating semantic analysis, are proposed for the approach. The proposed approach is applied in practice for predicting the price index movements of the top ten Chinese medicinal materials that receive the greatest media attention.

A set of experiments performed herein show that a predictive relationship exists between the news and the commodity market and that each of the three major methods improves the forecasting performance.

Because the field of Chinese medicinal materials lacks a corpus that can be used for sentiment analysis, the accuracy of a trained automatic sentiment classifier is lower than obtained by a manual method, which can cause the calculated convergence result to be inaccurate, thus affecting the final prediction model. The manual method of having people label news decreases the proposed method’s aspects of being intelligent and automatic.

Using the method proposed herein to predict the trends in Chinese medicinal materials prices helps farmers arrange a reasonable planting plan to pursue their best interests.

The method proposed herein to predict the trends in the prices of Chinese medicinal materials is conducive to the government arranging planned drug availabilities in order to avoid disasters in which herbs are looted.

The produced prediction result is meaningful in supporting farmers and investors to make better decisions in growing and trading Chinese medicinal material, which leads to financial returns on investments and the avoidance of severe losses.

Retail investors are prone to be affected by information dissemination in social media with the rapid development of Web 2.0. The purpose of this study is to recognize the factors that may impact users' retweet behavior, namely information dissemination in the online financial community, through machine learning techniques.

This paper crawled data from the Chinese online financial community (Xueqiu.com) and extracted author-related, content-related, situation-related, stock-related and stock market-related features from the dataset. The best information dissemination prediction model based on these features was determined by evaluating five classifiers with various performance metrics, and the predictability of different feature groups was tested.

Five prevalent classifiers were evaluated with various performance metrics and the random forest classifier was proven to be the best retweet prediction model in the authors’ experiments. Moreover, the predictability of author-related, content-related and market-related features was illustrated to be relatively better than that of the other two feature groups. Several particularly important features, such as the author's followers and the rise and fall of the stock index, were recognized in this paper at last.

This study contributes to in-depth research on information dissemination in the financial domain. The findings of this study have important practical implications for government regulators to supervise public opinion in the financial market.

While it has been claimed in many empirical studies that the political futures market can forecast better than the polls, it is unclear upon which our forecast should be based. Standard practice seems to suggest the use of the closing price of the market, as a reflection of the continuous process of information revealing and aggregation, but we are unsure that this practice applies to thin markets. In this chapter, we propose a number of reconstructions of the price series and use the closing price based on these reconstructed series as the forecast. We then test these ideas by comparing their forecasting performance with the closing price of the original series. It is found that forecasting accuracy can be gained if we use the closing price based on the smoothing series rather than the original series. However, there is no clear advantage by either using more sophisticated smoothing techniques, such as wavelets, or using external information, such as trading volume and duration time. The results show that the median, the simplest smoothing technique, performs rather well when compared with all complications.

The purpose of this paper is to present a new quantum‐inspired evolutionary hybrid intelligent (QIEHI) approach, in order to overcome the random walk dilemma for stock market prediction.

The proposed QIEHI method is inspired by the Takens' theorem and performs a quantum‐inspired evolutionary search for the minimum necessary dimension (time lags) embedded in the problem for determining the characteristic phase space that generates the financial time series phenomenon. The approach presented in this paper consists of a quantum‐inspired intelligent model composed of an artificial neural network (ANN) with a modified quantum‐inspired evolutionary algorithm (MQIEA), which is able to evolve the complete ANN architecture and parameters (pruning process), the ANN training algorithm (used to further improve the ANN parameters supplied by the MQIEA), and the most suitable time lags, to better describe the time series phenomenon.

This paper finds that, initially, the proposed QIEHI method chooses the better prediction model, then it performs a behavioral statistical test to adjust time phase distortions that appear in financial time series. Also, an experimental analysis is conducted with the proposed approach using six real‐word stock market times series, and the obtained results are discussed and compared, according to a group of relevant performance metrics, to results found with multilayer perceptron networks and the previously introduced time‐delay added evolutionary forecasting method.

The paper usefully demonstrates how the proposed QIEHI method chooses the best prediction model for the times series representation and performs a behavioral statistical test to adjust time phase distortions that frequently appear in financial time series.

The prediction of stock market (SM) indices is a fascinating task. An in-depth analysis in this field can provide valuable information to investors, traders and policy makers in attractive SMs. This article aims to apply a correlation feature selection model to identify important technical indicators (TIs), which are combined with multiple deep learning (DL) algorithms for forecasting SM indices.

The methodology involves using a correlation feature selection model to select the most relevant features. These features are then used to predict the fluctuations of six markets using various DL algorithms, and the results are compared with predictions made using all features by using a range of performance measures.

The experimental results show that the combination of TIs selected through correlation and Artificial Neural Network (ANN) provides good results in the MADEX market. The combination of selected indicators and Convolutional Neural Network (CNN) in the NASDAQ 100 market outperforms all other combinations of variables and models. In other markets, the combination of all variables with ANN provides the best results.

This article makes several significant contributions, including the use of a correlation feature selection model to select pertinent variables, comparison between multiple DL algorithms (ANN, CNN and Long-Short-Term Memory (LSTM)), combining selected variables with algorithms to improve predictions, evaluation of the suggested model on six datasets (MASI, MADEX, FTSE 100, SP500, NASDAQ 100 and EGX 30) and application of various performance measures (Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error(RMSE), Mean Squared Logarithmic Error (MSLE) and Root Mean Squared Logarithmic Error (RMSLE)).