Search results

For the case of many content features, This paper aims to investigate which content features in video and text ads more contribute to accurately predicting the success of crowdfunding by comparing prediction models.

With 1,368 features extracted from 15,195 Kickstarter campaigns in the USA, the authors compare base models such as logistic regression (LR) with tree-based homogeneous ensembles such as eXtreme gradient boosting (XGBoost) and heterogeneous ensembles such as XGBoost + LR.

XGBoost shows higher prediction accuracy than LR (82% vs 69%), in contrast to the findings of a previous relevant study. Regarding important content features, humans (e.g. founders) are more important than visual objects (e.g. products). In both spoken and written language, words related to experience (e.g. eat) or perception (e.g. hear) are more important than cognitive (e.g. causation) words. In addition, a focus on the future is more important than a present or past time orientation. Speech aids (see and compare) to complement visual content are also effective and positive tone matters in speech.

This research makes theoretical contributions by finding more important visuals (human) and language features (experience, perception and future time). Also, in a multimodal context, complementary cues (e.g. speech aids) across different modalities help. Furthermore, the noncontent parts of speech such as positive “tone” or pace of speech are important.

Founders are encouraged to assess and revise the content of their video or text ads as well as their basic campaign features (e.g. goal, duration and reward) before they launch their campaigns. Next, overly complex ensembles may suffer from overfitting problems. In practice, model validation using unseen data is recommended.

Rather than reducing the number of content feature dimensions (Kaminski and Hopp, 2020), by enabling advanced prediction models to accommodate many contents features, prediction accuracy rises substantially.

The study seeks to identify the impact of variables in a deep learning-based bankruptcy prediction model, which has achieved superior performance to other prediction models but cannot easily interpret hidden processes.

This study developed three LSTM-RNN–based models that predicted the probability of bankruptcy before 1, 2 and 3 years using financial, the construction market and macroeconomic variables as input variables. Then, the impacts of the input variables that affected prediction accuracy in each model were identified by using Shapley value and compared among the three models. This study also investigated the prediction accuracy using variants of input variables grouped sequentially by high-impact ranking.

The results showed that the prediction accuracies were largely impacted by “housing starts” in all models. As the prediction period increased, the effects of macroeconomic variables on prediction accuracy increased, whereas the impact of “return on assets” on prediction accuracy decreased. It also found that the “current ratio” and “debt ratio” significantly influenced the prediction accuracies in all models. Also, the results revealed that similar prediction accuracies could be achieved using only 8, 10, and 10 variables out of a total of 18 variables for the 1-, 2-, and 3-year prediction models, respectively.

This study provides a Shapley value-based approach to identify how each input variable in a deep-learning bankruptcy prediction model. The findings of this study can not only assist in obtaining better insights into the underlying concept of bankruptcy but also use to select variables by removing those identified as less significant.

The purpose of this paper is to establish a random simulation method to compare the forecasting performance between grey prediction models, and between grey model and other kinds of prediction models. Then, the different performance of three grey models and linear regression prediction model is studied, based on the proposed method.

A random simulation method was proposed to test the modelling accuracy of grey prediction model. This method was enlightened by Monte Carlo simulation method. It regarded a class of sequences as population, and selected a large sample from population though random sampling. Then, sample sequences were modeled by grey prediction model. Through modeling error calculation, the average error of grey model for the sample was obtained. Finally, the grey model accuracy for this kind of problem was acquired by statistical inference testing model. Through the statistical significant test method, the modeling accuracy of grey models for the same problem can be compared. Also, accuracy difference between grey prediction model and regression analysis, support vector machine, neural network, and other forecasting methods can be also compared.

Though random simulation experiments, the following conclusion was obtained. First, grey model can be applied to the long sequence whose growth rate was less than 20 per cent, and the short sequence whose growth rate was less than 50 per cent. Second, GM(1,1) cannot be applied to a long sequence with high growth. Third, growth rate was a more important factor than growth length on modeling accuracy of GM(1,1). Fourth, when the sequence length was short, accuracy of GM(1,1) model was higher than linear regression. While the length of the sequence was more than 15, and the growth rate in [0‐10 per cent], two kinds of modeling error was not significantly different.

The method proposed in the paper can be used to compare the performance of different prediction models, and to select appropriate model for a prediction problem.

The paper succeeded in establishing an accuracy test method for grey models and other prediction models. It will standardize the grey modelling and contribute to application of grey models.

This paper aims to provide a predictive framework of customer churn through six stages for accurate prediction and preventing customer churn in the field of business.

The six stages are as follows: first, collection of customer behavioral data and preparation of the data; second, the formation of derived variables and selection of influential variables, using a method of discriminant analysis; third, selection of training and testing data and reviewing their proportion; fourth, the development of prediction models using simple, bagging and boosting versions of supervised machine learning; fifth, comparison of churn prediction models based on different versions of machine-learning methods and selected variables; and sixth, providing appropriate strategies based on the proposed model.

According to the results, five variables, the number of items, reception of returned items, the discount, the distribution time and the prize beside the recency, frequency and monetary (RFM) variables (RFMITSDP), were chosen as the best predictor variables. The proposed model with accuracy of 97.92 per cent, in comparison to RFM, had much better performance in churn prediction and among the supervised machine learning methods, artificial neural network (ANN) had the highest accuracy, and decision trees (DT) was the least accurate one. The results show the substantially superiority of boosting versions in prediction compared with simple and bagging models.

The period of the available data was limited to two years. The research data were limited to only one grocery store whereby it may not be applicable to other industries; therefore, generalizing the results to other business centers should be used with caution.

Business owners must try to enforce a clear rule to provide a prize for a certain number of purchased items. Of course, the prize can be something other than the purchased item. Business owners must accept the items returned by the customers for any reasons, and the conditions for accepting returned items and the deadline for accepting the returned items must be clearly communicated to the customers. Store owners must consider a discount for a certain amount of purchase from the store. They have to use an exponential rule to increase the discount when the amount of purchase is increased to encourage customers for more purchase. The managers of large stores must try to quickly deliver the ordered items, and they should use equipped and new transporting vehicles and skilled and friendly workforce for delivering the items. It is recommended that the types of services, the rules for prizes, the discount, the rules for accepting the returned items and the method of distributing the items must be prepared and shown in the store for all the customers to see. The special services and reward rules of the store must be communicated to the customers using new media such as social networks. To predict the customer behaviors based on the data, the future researchers should use the boosting method because it increases efficiency and accuracy of prediction. It is recommended that for predicting the customer behaviors, particularly their churning status, the ANN method be used. To extract and select the important and effective variables influencing customer behaviors, the discriminant analysis method can be used which is a very accurate and powerful method for predicting the classes of the customers.

The current study tries to fill this gap by considering five basic and important variables besides RFM in stores, i.e. prize, discount, accepting returns, delay in distribution and the number of items, so that the business owners can understand the role services such as prizes, discount, distribution and accepting returns play in retraining the customers and preventing them from churning. Another innovation of the current study is the comparison of machine-learning methods with their boosting and bagging versions, especially considering the fact that previous studies do not consider the bagging method. The other reason for the study is the conflicting results regarding the superiority of machine-learning methods in a more accurate prediction of customer behaviors, including churning. For example, some studies introduce ANN (Huang et al., 2010; Hung and Wang, 2004; Keramati et al., 2014; Runge et al., 2014), some introduce support vector machine ( Guo-en and Wei-dong, 2008; Vafeiadis et al., 2015; Yu et al., 2011) and some introduce DT (Freund and Schapire, 1996; Qureshi et al., 2013; Umayaparvathi and Iyakutti, 2012) as the best predictor, confusing the users of the results of these studies regarding the best prediction method. The current study identifies the best prediction method specifically in the field of store businesses for researchers and the owners. Moreover, another innovation of the current study is using discriminant analysis for selecting and filtering variables which are important and effective in predicting churners and non-churners, which is not used in previous studies. Therefore, the current study is unique considering the used variables, the method of comparing their accuracy and the method of selecting effective variables.

To improve the speed and accuracy of turbine blade film cooling design process, the most advanced deep learning models were introduced into this study to investigate the most suitable define for prediction work. This paper aims to create a generative surrogate model that can be applied on multi-objective optimization problems.

The latest backbone in the field of computer vision (Swin-Transformer, 2021) was introduced and improved as the surrogate function for prediction of the multi-physics field distribution (film cooling effectiveness, pressure, density and velocity). The basic samples were generated by Latin hypercube sampling method and the numerical method adopt for the calculation was validated experimentally at first. The training and testing samples were calculated at experimental conditions. At last, the surrogate model predicted results were verified by experiment in a linear cascade.

The results indicated that comparing with the Multi-Scale Pix2Pix Model, the Swin-Transformer U-Net model presented higher accuracy and computing speed on the prediction of contour results. The computation time for each step of the Swin-Transformer U-Net model is one-third of the original model, especially in the case of multi-physics field prediction. The correlation index reached more than 99.2% and the first-order error was lower than 0.3% for multi-physics field. The predictions of the data-driven surrogate model are consistent with the predictions of the computational fluid dynamics results, and both are very close to the experimental results. The application of the Swin-Transformer model on enlarging the different structure samples will reduce the cost of numerical calculations as well as experiments.

The number of U-Net layers and sample scales has a proper relationship according to equation (8). Too many layers of U-Net will lead to unnecessary nonlinear variation, whereas too few layers will lead to insufficient feature extraction. In the case of Swin-Transformer U-Net model, incorrect number of U-Net layer will reduce the prediction accuracy. The multi-scale Pix2Pix model owns higher accuracy in predicting a single physical field, but the calculation speed is too slow. The Swin-Transformer model is fast in prediction and training (nearly three times faster than multi Pix2Pix model), but the predicted contours have more noise. The neural network predicted results and numerical calculations are consistent with the experimental distribution.

This paper creates a generative surrogate model that can be applied on multi-objective optimization problems. The generative adversarial networks using new backbone is chosen to adjust the output from single contour to multi-physics fields, which will generate more results simultaneously than traditional surrogate models and reduce the time-cost. And it is more applicable to multi-objective spatial optimization algorithms. The Swin-Transformer surrogate model is three times faster to computation speed than the Multi Pix2Pix model. In the prediction results of multi-physics fields, the prediction results of the Swin-Transformer model are more accurate.

The purpose of this paper is to improve the information quality of bankruptcy prediction models proposed in the literature by building prediction intervals around the point estimates generated by these models and to determine if the use of the prediction intervals in conjunction with the point estimated yields an improvement in predictive accuracy over traditional models. The authors calculated the point estimates and prediction intervals for a sample of firms from 1991 to 2008. The point estimates and prediction intervals were used in concert to classify firms as bankrupt or non-bankrupt. The accuracy of the tested technique was compared to that of a traditional bankruptcy prediction model. The results indicate that the use of upper and lower bounds in concert with the point estimates yield an improvement in the predictive ability of bankruptcy prediction models. The improvements in overall prediction accuracy and non-bankrupt firm prediction accuracy are statistically significant at the 0.01 level. The authors present a technique that (1) provides a more complete picture of the firm’s status, (2) is derived from multiple forms of evidence, (3) uses a predictive interval technique that is easily repeated, (4) can be generated in a timely manner, (5) can be applied to other bankruptcy prediction models in the literature, and (6) is statistically significantly more accurate than traditional point estimate techniques. The current research is the first known study to use the combination of point estimates and prediction intervals to in bankruptcy prediction.

The paper attempts to establish GM(1,1) grey prediction model group for the top three Olympic track and field sports performance, and to predict the 30th London Olympic track and field results and its tendency using grey systems theory.

Athletics sports achievements are influenced by many factors, such as the physical quality, athletes individual growth cycle, and injuring or retirement of excellent athletes, the outstanding performance of some athletes, the using of high‐tech sports training instrument, the implementation plan of scientific training guidance, the introduction of advanced technology, facilities and improvement, and so on. Those aspects can make the match result uncertain, which are running in a uncertain and continually changing environment, so sports achievements have obviously grey features. Combined with grey modeling methods, and aimed at the top three Olympic track and field sports performance, this paper established GM (1,1) grey prediction model group and analysed the trend of Olympic track and field. And in the end of the paper, the 30th Olympic men's and women's the top three athletic achievements prediction intervals are also predicted.

The results show that forecasting model group has high‐precision. In the 46 champions prediction models, three models have the forecast accuracy of 100 percent; 27 models' forecast accuracy are greater than 99.5 percent, and the rest of the models forecast accuracy are greater than 98.58 percent. In the 46 silver medalists prediction models, five models have the forecast accuracy of 100 percent; 33 models' forecast accuracy are greater than 99.5 percent and the rest of the models' forecast accuracy is greater than 98.48 percent. In the 46 bronze medalist prediction models, four models have the forecast accuracy of 100 percent; 25 models' forecast accuracy is greater than 99.5 percent and the rest of the models forecast accuracy is greater than 98.76 percent. The essay deeply analyzes the top three achievements' trend of Olympic Games Track and field. In the end, the paper predicts the 30th Olympic track and field results.

The method exposed in the paper can be used for the short‐term or long‐term prediction of sports scores metering in international competition (such as track and field, swimming, rowing, etc.), and also for personal athletic performance prediction.

The paper succeeds in realising both grey prediction model group for the top three Olympic track and field performance in all projects, and prediction of the 30th London Olympic track and field results by using the newest developed theories: grey systems theory.

Many factors affect the emergence and development of crop diseases and insect pests. Traditional methods for investigating this subject are often difficult to employ and produce limited data with considerable uncertainty. The purpose of this paper is to predict the annual degree of cotton spider mite infestations by employing grey theory.

The authors established a GM(1,1) model to forecast mite infestation degree based on the analysis of historical data. To improve the prediction accuracy, the authors modified the grey model using Markov chain and BP neural network analyses. The prediction accuracy of the GM(1,1), Grey-Markov chain, and Grey-BP neural network models was 84.31, 94.76, and 96.84 per cent, respectively.

Compared with the single grey forecast model, both the Grey-Markov chain model and the Grey-BP neural network model had higher forecast accuracy, and the accuracy of the latter was highest. The improved grey model can be used to predict the degree of cotton spider mite infestations with high accuracy and overcomes the shortcomings of traditional forecasting methods.

The two new models were used to estimate mite infestation degree in 2015 and 2016. The Grey-Markov chain model yielded respective values of 1.27 and 1.15, whereas the Grey-BP neural network model yielded values 1.4 and 1.68; the actual values were 1.5 and 1.8.

The improved grey model can be used for medium- and long-term predictions of the occurrence of cotton spider mites and overcomes problems caused by data singularity and fluctuation. This research method can provide a reference for the prediction of similar diseases.

The main aim of the study is to identify some critical microeconomic determinants of financial distress and to design a parsimonious distress prediction model for an emerging economy like India. In doing so, the authors also attempt to compare the forecasting accuracy of alternative distress prediction techniques.

In this study, the authors use two alternatives accounting information-based definitions of financial distress to construct a measure of financial distress. The authors then use the binomial logit model and two other popular machine learning–based models, namely artificial neural network and support vector machine, to compare the distress prediction accuracy rate of these alternative techniques for the Indian corporate sector.

The study’s empirical results suggest that five financial ratios, namely return on capital employed, cash flows to total liability, asset turnover ratio, fixed assets to total assets, debt to equity ratio and a measure of firm size (log total assets), play a highly significant role in distress prediction. The study’s findings suggest that machine learning-based models, namely support vector machine (SVM) and artificial neural network (ANN), are superior in terms of their prediction accuracy compared to the simple binomial logit model. Results also suggest that one-year-ahead forecasts are relatively better than the two-year-ahead forecasts.

The findings of the study have some important practical implications for creditors, policymakers, regulators and other stakeholders. First, rather than monitoring and collecting information on a list of predictor variables, only six most important accounting ratios may be monitored to track the transition of a healthy firm into financial distress. Second, our six-factor model can be used to devise a sound early warning system for corporate financial distress. Three, machine learning–based distress prediction models have prediction accuracy superiority over the commonly used time series model in the available literature for distress prediction involving a binary dependent variable.

This study is one of the first comprehensive attempts to investigate and design a parsimonious distress prediction model for the emerging Indian economy which is currently facing high levels of corporate financial distress. Unlike the previous studies, the authors use two different accounting information-based measures of financial distress in order to identify an effective way of measuring financial distress. Some of the determinants of financial distress identified in this study are different from the popular distress prediction models used in the literature. Our distress prediction model can be useful for the other emerging markets for distress prediction.

The purpose of this paper is to develop a common remaining shelf life prediction model that is generally applicable for postharvest table grape using an optimized radial basis function (RBF) neural network to achieve more accurate prediction than the current shelf life (SL) prediction methods.

First, the final indicators (storage temperature, relative humidity, sensory average score, peel hardness, soluble solids content, weight loss rate, rotting rate, fragmentation rate and color difference) affecting SL were determined by the correlation and significance analysis. Then using the analytic hierarchy process (AHP) to calculate the weight of each indicator and determine the end of SL under different storage conditions. Subsequently, the structure of the RBF network redesigned was 9-11-1. Ultimately, the membership degree of Fuzzy clustering (fuzzy c-means) was adopted to optimize the center and width of the RBF network by using the training data.

The results show that this method has the highest prediction accuracy compared to the current the kinetic–Arrhenius model, back propagation (BP) network and RBF network. The maximum absolute error is 1.877, the maximum relative error (RE) is 0.184, and the adjusted R² is 0.911. The prediction accuracy of the kinetic–Arrhenius model is the worst. The RBF network has a better prediction accuracy than the BP network. For robustness, the adjusted R² are 0.853 and 0.886 of Italian grape and Red Globe grape, respectively, and the fitting degree are the highest among all methods, which proves that the optimized method is applicable for accurate SL prediction of different table grape varieties.

This study not only provides a new way for the prediction of SL of different grape varieties, but also provides a reference for the quality and safety management of table grape during storage. Maybe it has a further research significance for the application of RBF neural network in the SL prediction of other fresh foods.