Pub Date : 2019-05-04DOI: 10.1109/CIFEr.2019.8759117
Yue Liu, Adam Ghandar, G. Theodoropoulos
As countries develop digital financial infrastructure, a wide range of economic activities expand and grow in importance: from personal loans, to the rapidly developing networked microfinance industry, to mobile telephone services and real estate transactions and so on. Personal credit is also a foundation of trust for facilitation of integrated societal transactions more generally. In emerging markets there is, however, a gap between the requirement for establishing a credit or trust rating and the lack of a credit record. The development of methodologies for greater financial integration of growing economies has the potential to have a significant impact on increasing the GDP of developing economies (4-12% according to a recent McKinsey Global Institute report). In this paper, we develop and test a methodology for feature selection and test its in standard datasets from large institutions in mature market economies, and a recent dataset which illustrates characteristics of emerging markets. The results show performance in classification can be maintained while runtime can be reduced when using a GA for feature selection in a range of machine learning techniques.
{"title":"A Metaheuristic Strategy for Feature Selection Problems: Application to Credit Risk Evaluation in Emerging Markets","authors":"Yue Liu, Adam Ghandar, G. Theodoropoulos","doi":"10.1109/CIFEr.2019.8759117","DOIUrl":"https://doi.org/10.1109/CIFEr.2019.8759117","url":null,"abstract":"As countries develop digital financial infrastructure, a wide range of economic activities expand and grow in importance: from personal loans, to the rapidly developing networked microfinance industry, to mobile telephone services and real estate transactions and so on. Personal credit is also a foundation of trust for facilitation of integrated societal transactions more generally. In emerging markets there is, however, a gap between the requirement for establishing a credit or trust rating and the lack of a credit record. The development of methodologies for greater financial integration of growing economies has the potential to have a significant impact on increasing the GDP of developing economies (4-12% according to a recent McKinsey Global Institute report). In this paper, we develop and test a methodology for feature selection and test its in standard datasets from large institutions in mature market economies, and a recent dataset which illustrates characteristics of emerging markets. The results show performance in classification can be maintained while runtime can be reduced when using a GA for feature selection in a range of machine learning techniques.","PeriodicalId":368382,"journal":{"name":"2019 IEEE Conference on Computational Intelligence for Financial Engineering & Economics (CIFEr)","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-05-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129446495","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-05-04DOI: 10.1109/CIFEr.2019.8759064
Yifan Li, Xuan Wang, Z. L. Jiang, Shuhan Qi, Xinhui Liu, Qian Chen
How to track the target robustly is a challenging task in the field of computer vision. Occlusion as one of the most difficult problems, occurs due to the information lost when three-dimensional subjects are projected in two-dimensional interface, therefore, the 2D or 3D tracking algorithms which adopted depth information that expects to rely on three-dimensional special structure to resolve these problems and made somewhat progress. The 2D tracking algorithm is not efficient in fully using depth information, and the 3D tracking method is not robust because of the lack of mature 3D feature extraction method, which fairly restricts the actual tracking effect. Responding to above questions, we propose an adoption of adaptive quantified depth information, establish an adaptive hierarchical structure according to various scenarios. Hierarchical structure can filter the foreground and background information to reduce the interference in tracking, at the same time simplify the use of the depth information. Combined with kernel correlation filter tracking method, we design the algorithm using 2D apparent model under the spatial structures, which is efficient to deal with the problems of occlusion and the change of target scale, and prove its effectiveness on Princeton Tracking Dataset.
{"title":"RGB-D tracker under Hierarchical structure","authors":"Yifan Li, Xuan Wang, Z. L. Jiang, Shuhan Qi, Xinhui Liu, Qian Chen","doi":"10.1109/CIFEr.2019.8759064","DOIUrl":"https://doi.org/10.1109/CIFEr.2019.8759064","url":null,"abstract":"How to track the target robustly is a challenging task in the field of computer vision. Occlusion as one of the most difficult problems, occurs due to the information lost when three-dimensional subjects are projected in two-dimensional interface, therefore, the 2D or 3D tracking algorithms which adopted depth information that expects to rely on three-dimensional special structure to resolve these problems and made somewhat progress. The 2D tracking algorithm is not efficient in fully using depth information, and the 3D tracking method is not robust because of the lack of mature 3D feature extraction method, which fairly restricts the actual tracking effect. Responding to above questions, we propose an adoption of adaptive quantified depth information, establish an adaptive hierarchical structure according to various scenarios. Hierarchical structure can filter the foreground and background information to reduce the interference in tracking, at the same time simplify the use of the depth information. Combined with kernel correlation filter tracking method, we design the algorithm using 2D apparent model under the spatial structures, which is efficient to deal with the problems of occlusion and the change of target scale, and prove its effectiveness on Princeton Tracking Dataset.","PeriodicalId":368382,"journal":{"name":"2019 IEEE Conference on Computational Intelligence for Financial Engineering & Economics (CIFEr)","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-05-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116579217","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-05-04DOI: 10.1109/CIFEr.2019.8759056
Weiwei Zhang, Chao Zhou
Portfolio selection with proportional transaction cost is a singular stochastic control problem that has been widely discussed. In this paper, we propose a deep learning based numerical scheme to solve transaction cost problems, and compare its effectiveness with a penalty partial differential equation (PDE) method. We further extend it to multi-asset cases which existing numerical methods can not be applied to due to the curse of dimensionality. Deep learning algorithm directly approximates the optimal trading strategies by a feedforward neural network at each discrete time. It is observed that deep learning approach can achieve satisfying performance to characterize optimal buy and sell boundaries and thus value function.
{"title":"Deep Learning Algorithm to solve Portfolio Management with Proportional Transaction Cost","authors":"Weiwei Zhang, Chao Zhou","doi":"10.1109/CIFEr.2019.8759056","DOIUrl":"https://doi.org/10.1109/CIFEr.2019.8759056","url":null,"abstract":"Portfolio selection with proportional transaction cost is a singular stochastic control problem that has been widely discussed. In this paper, we propose a deep learning based numerical scheme to solve transaction cost problems, and compare its effectiveness with a penalty partial differential equation (PDE) method. We further extend it to multi-asset cases which existing numerical methods can not be applied to due to the curse of dimensionality. Deep learning algorithm directly approximates the optimal trading strategies by a feedforward neural network at each discrete time. It is observed that deep learning approach can achieve satisfying performance to characterize optimal buy and sell boundaries and thus value function.","PeriodicalId":368382,"journal":{"name":"2019 IEEE Conference on Computational Intelligence for Financial Engineering & Economics (CIFEr)","volume":"36 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-05-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127749093","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-05-04DOI: 10.1109/CIFEr.2019.8759109
Le Lv, Jianbo Cheng, Nanbo Peng, Min Fan, Dongbin Zhao, Jianhong Zhang
Many practical problems can be formulated as graph-based semi-supervised classification problems. For example, online finance anti-fraud. Recently, many researchers attempt using deep learning methods to solve such problems. In this paper, we propose a novel neural network architecture to perform semi-supervised classification on graph-structured data. We improve the graph convolutional network (GCN) by replacing the graph convolution matrix with auto-encoder module. The proposed neural network is trained by a multi-task objective function. Except the classification task, we train the auto-encoder module to reconstruct the graph convolution matrix. It can be seen as an adaptive spectral convolution on graph. It can increase the depth of neural network without causing over-smooth effect. Additionally, the introduction of reconstruction task can mitigate the cold-start problem. Even the graph topological structure is extreme sparse, our method can learn expressive latent features for vertices. The experimental results show that our method can achieve the state of art performance.
{"title":"Auto-encoder based Graph Convolutional Networks for Online Financial Anti-fraud","authors":"Le Lv, Jianbo Cheng, Nanbo Peng, Min Fan, Dongbin Zhao, Jianhong Zhang","doi":"10.1109/CIFEr.2019.8759109","DOIUrl":"https://doi.org/10.1109/CIFEr.2019.8759109","url":null,"abstract":"Many practical problems can be formulated as graph-based semi-supervised classification problems. For example, online finance anti-fraud. Recently, many researchers attempt using deep learning methods to solve such problems. In this paper, we propose a novel neural network architecture to perform semi-supervised classification on graph-structured data. We improve the graph convolutional network (GCN) by replacing the graph convolution matrix with auto-encoder module. The proposed neural network is trained by a multi-task objective function. Except the classification task, we train the auto-encoder module to reconstruct the graph convolution matrix. It can be seen as an adaptive spectral convolution on graph. It can increase the depth of neural network without causing over-smooth effect. Additionally, the introduction of reconstruction task can mitigate the cold-start problem. Even the graph topological structure is extreme sparse, our method can learn expressive latent features for vertices. The experimental results show that our method can achieve the state of art performance.","PeriodicalId":368382,"journal":{"name":"2019 IEEE Conference on Computational Intelligence for Financial Engineering & Economics (CIFEr)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-05-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130877295","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-05-01DOI: 10.1109/CIFEr.2019.8759057
Atsuki Nakayama, K. Izumi, Hiroki Sakaji, Hiroyasu Matsushima, T. Shimada, Kenta Yamada
Predicting the price movements of stocks based on deep learning and high-frequency data has been studied intensively in recent years. Especially, limit order book which describes the supply-demand balance of the market is used as features of a neural network. However, these methods do not utilize the properties of market orders. On the other hand, this study encodes information of time and prices of orders into images. This encoding method can take advantage of these properties. Then, we apply machine learning methods, convolutional neural network (CNN) and logistic regression (LR), to order-based features to predict the direction of short-term price movements. The results show that the execution has the highest prediction power than the order and cancellation information. Moreover, the difference between CNN and LR are small and depends on kinds of stocks.
{"title":"Short-term Stock Price Prediction by Analysis of Order Pattern Images","authors":"Atsuki Nakayama, K. Izumi, Hiroki Sakaji, Hiroyasu Matsushima, T. Shimada, Kenta Yamada","doi":"10.1109/CIFEr.2019.8759057","DOIUrl":"https://doi.org/10.1109/CIFEr.2019.8759057","url":null,"abstract":"Predicting the price movements of stocks based on deep learning and high-frequency data has been studied intensively in recent years. Especially, limit order book which describes the supply-demand balance of the market is used as features of a neural network. However, these methods do not utilize the properties of market orders. On the other hand, this study encodes information of time and prices of orders into images. This encoding method can take advantage of these properties. Then, we apply machine learning methods, convolutional neural network (CNN) and logistic regression (LR), to order-based features to predict the direction of short-term price movements. The results show that the execution has the highest prediction power than the order and cancellation information. Moreover, the difference between CNN and LR are small and depends on kinds of stocks.","PeriodicalId":368382,"journal":{"name":"2019 IEEE Conference on Computational Intelligence for Financial Engineering & Economics (CIFEr)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117058351","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-05-01DOI: 10.1109/CIFEr.2019.8759119
Kyoto Yono, K. Izumi, Hiroki Sakaji, Hiroyasu Matsushima, T. Shimada
For financial market participants, the current focused topic (Brexit, Federal Reserve Interest-Rate, U.S. and China trade war, etc.) and its sentiments (whether it is Risk-On or Risk-Off) is very important to decide investment strategies. In this study, we proposed extended topic model called supervised Joint Sentiment-Topic model (sJST) which using not only text data but also numeric data as a supervised signal to extract current focused topic and it's sentiment of market. By using the topic and sentiment weight of the market as a features, we apply several machine learning models to predict foreign exchange market price movement. Comparing the average accuracy over 32 currency pairs and prediction models, the result using sJST weight as features achieved 1.52% better performance than the results which use only historical prices as features. Furthermore, comparing the results limited to specific currency pairs which is difficult to predict when using only historical prices as features, the result using sJST weight as features achieved 3.64% better accuracy than the result which use only historical prices as features.
{"title":"Extraction of Focused Topic and Sentiment of Financial Market by using Supervised Topic Model for Price Movement Prediction","authors":"Kyoto Yono, K. Izumi, Hiroki Sakaji, Hiroyasu Matsushima, T. Shimada","doi":"10.1109/CIFEr.2019.8759119","DOIUrl":"https://doi.org/10.1109/CIFEr.2019.8759119","url":null,"abstract":"For financial market participants, the current focused topic (Brexit, Federal Reserve Interest-Rate, U.S. and China trade war, etc.) and its sentiments (whether it is Risk-On or Risk-Off) is very important to decide investment strategies. In this study, we proposed extended topic model called supervised Joint Sentiment-Topic model (sJST) which using not only text data but also numeric data as a supervised signal to extract current focused topic and it's sentiment of market. By using the topic and sentiment weight of the market as a features, we apply several machine learning models to predict foreign exchange market price movement. Comparing the average accuracy over 32 currency pairs and prediction models, the result using sJST weight as features achieved 1.52% better performance than the results which use only historical prices as features. Furthermore, comparing the results limited to specific currency pairs which is difficult to predict when using only historical prices as features, the result using sJST weight as features achieved 3.64% better accuracy than the result which use only historical prices as features.","PeriodicalId":368382,"journal":{"name":"2019 IEEE Conference on Computational Intelligence for Financial Engineering & Economics (CIFEr)","volume":"83 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132854843","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-05-01DOI: 10.1109/CIFEr.2019.8759106
Reza Refaei Afshar, Yingqian Zhang, M. Firat, U. Kaymak
Online advertising is one of the most important sources of income for many online publishers. The process is as easy as placing slots in the website and selling those slots in real time bidding auctions. Since websites load in few milliseconds, the bidding and selling process should not take too much time. Sellers or publishers of advertisements aim to maximize the revenue obtained through online advertising. In this paper, we propose a method to select the most profitable ad network for each ad request that is built upon our previous work [1]. The proposed method consists of two parts: a prediction model and a reinforcement learning modeling. We test two strategies of selecting ad network orderings. The first strategy uses the developed prediction model to greedily choose the network with the highest expected revenue. The second strategy is a two-step approach, where a reinforcement learning method is used to improve the revenue estimation of the prediction model. Using real AD auction data, we show that the ad network ordering obtained from the second strategy returns much higher revenue than the first strategy.
{"title":"A Decision Support Method to Increase the Revenue of Ad Publishers in Waterfall Strategy","authors":"Reza Refaei Afshar, Yingqian Zhang, M. Firat, U. Kaymak","doi":"10.1109/CIFEr.2019.8759106","DOIUrl":"https://doi.org/10.1109/CIFEr.2019.8759106","url":null,"abstract":"Online advertising is one of the most important sources of income for many online publishers. The process is as easy as placing slots in the website and selling those slots in real time bidding auctions. Since websites load in few milliseconds, the bidding and selling process should not take too much time. Sellers or publishers of advertisements aim to maximize the revenue obtained through online advertising. In this paper, we propose a method to select the most profitable ad network for each ad request that is built upon our previous work [1]. The proposed method consists of two parts: a prediction model and a reinforcement learning modeling. We test two strategies of selecting ad network orderings. The first strategy uses the developed prediction model to greedily choose the network with the highest expected revenue. The second strategy is a two-step approach, where a reinforcement learning method is used to improve the revenue estimation of the prediction model. Using real AD auction data, we show that the ad network ordering obtained from the second strategy returns much higher revenue than the first strategy.","PeriodicalId":368382,"journal":{"name":"2019 IEEE Conference on Computational Intelligence for Financial Engineering & Economics (CIFEr)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121147556","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-05-01DOI: 10.1109/CIFEr.2019.8759111
Jun Chen, E. Tsang
In our previous work, we showed that regime changes in the market are retrospectively detectable using historic data in directional changes (DC). In this paper, we build on such results and show that DC indicators can be used for market tracking - using data up to the present - to understand what is going on in the market. In particular, we wanted to track the market to see whether the market is entering an abnormally volatile regime. The proposed approach used DC indicator values observed in the past to model the normal regime of a market (in which volatility is normal) or an abnormal regime (in which volatility is abnormally high). Given a particular value observed in the current market, we used a naive Bayes model to calculate independently two probabilities: one for the market being in the normal regime and one for it being in the abnormal regime. These two probabilities were combined to decide which regime the market was in, two decision rules were examined: a Simple Rule and a Stricter Rule. We used DJIA, FTSE 100 and S&P 500 data from 2007 to 2010 to build the Bayes model. The model was used to track the S&P 500 market from 2010 to 2012, which saw two spells of abnormal regimes, as identified by our previous work, with the benefit of hindsight. The tracking method presented in this paper, with either decision rule, managed to pick up both spells of regime changes accurately. The tracking signals could be useful to market participants. This study potentially lays the foundation of a practical financial early warning system.
{"title":"Tacking Regime Changes in the Markets","authors":"Jun Chen, E. Tsang","doi":"10.1109/CIFEr.2019.8759111","DOIUrl":"https://doi.org/10.1109/CIFEr.2019.8759111","url":null,"abstract":"In our previous work, we showed that regime changes in the market are retrospectively detectable using historic data in directional changes (DC). In this paper, we build on such results and show that DC indicators can be used for market tracking - using data up to the present - to understand what is going on in the market. In particular, we wanted to track the market to see whether the market is entering an abnormally volatile regime. The proposed approach used DC indicator values observed in the past to model the normal regime of a market (in which volatility is normal) or an abnormal regime (in which volatility is abnormally high). Given a particular value observed in the current market, we used a naive Bayes model to calculate independently two probabilities: one for the market being in the normal regime and one for it being in the abnormal regime. These two probabilities were combined to decide which regime the market was in, two decision rules were examined: a Simple Rule and a Stricter Rule. We used DJIA, FTSE 100 and S&P 500 data from 2007 to 2010 to build the Bayes model. The model was used to track the S&P 500 market from 2010 to 2012, which saw two spells of abnormal regimes, as identified by our previous work, with the benefit of hindsight. The tracking method presented in this paper, with either decision rule, managed to pick up both spells of regime changes accurately. The tracking signals could be useful to market participants. This study potentially lays the foundation of a practical financial early warning system.","PeriodicalId":368382,"journal":{"name":"2019 IEEE Conference on Computational Intelligence for Financial Engineering & Economics (CIFEr)","volume":"68 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116728946","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-05-01DOI: 10.1109/CIFEr.2019.8759112
Jia-Hao Syu, Mu-En Wu, Shin-Huah Lee, Jan-Ming Ho
Opening Range Breakout (ORB) is a fairly intraday trading strategy. We set the resistance and the support levels by the price in opening interval to follow the trend in the futures market. However, such kind of strategies is not profitable for most commodities in recent years in the changing market. In this paper, we attempt to improve the original ORB strategy by considering the effect of trends continuity on the event. We adjust the predetermined threshold for upper bound and lower bound. This strategy is called Threshold Adjusting ORB or TA_ORB. We implement this modified ORB strategy on the Taiwan Index Futures from 2008 to 2012. Compared with the original ORB strategy, we got 145.98% return in 2008 (bear market), 81.86% return in 2009 (bull market) and 32.25% annual return in 2008–2012 (five-year period) which are 4.0 times, 1.4 times, and 2.6 times more than original ORB, respectively. TA_ORB performs outstanding in large fluctuation, especially in the bear market. Performance can verify that the observations of TA_ORB improve the stability of the breakthrough signal, enhance the return, and reduce strategic risk. Further, we plan to use neural network to make more precise predictions and implement these strategies in different commodities.
{"title":"Modified ORB Strategies with Threshold Adjusting on Taiwan Futures Market","authors":"Jia-Hao Syu, Mu-En Wu, Shin-Huah Lee, Jan-Ming Ho","doi":"10.1109/CIFEr.2019.8759112","DOIUrl":"https://doi.org/10.1109/CIFEr.2019.8759112","url":null,"abstract":"Opening Range Breakout (ORB) is a fairly intraday trading strategy. We set the resistance and the support levels by the price in opening interval to follow the trend in the futures market. However, such kind of strategies is not profitable for most commodities in recent years in the changing market. In this paper, we attempt to improve the original ORB strategy by considering the effect of trends continuity on the event. We adjust the predetermined threshold for upper bound and lower bound. This strategy is called Threshold Adjusting ORB or TA_ORB. We implement this modified ORB strategy on the Taiwan Index Futures from 2008 to 2012. Compared with the original ORB strategy, we got 145.98% return in 2008 (bear market), 81.86% return in 2009 (bull market) and 32.25% annual return in 2008–2012 (five-year period) which are 4.0 times, 1.4 times, and 2.6 times more than original ORB, respectively. TA_ORB performs outstanding in large fluctuation, especially in the bear market. Performance can verify that the observations of TA_ORB improve the stability of the breakthrough signal, enhance the return, and reduce strategic risk. Further, we plan to use neural network to make more precise predictions and implement these strategies in different commodities.","PeriodicalId":368382,"journal":{"name":"2019 IEEE Conference on Computational Intelligence for Financial Engineering & Economics (CIFEr)","volume":"52 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133392789","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-05-01DOI: 10.1109/CIFEr.2019.8759061
Yu-Fei Lin, Yeong-Luh Ueng, W. Chung, Tzu-Ming Huang
By knowing the future price range, which is the difference between the closing price and the opening price, we can calculate the long or short positions in advance. This paper presents a Recurrent Neural Network (RNN) based approach to forecast the price range. Compared to other methods based on machine learning, our method puts greater focus on the characteristics of the stock data, such as the zero-crossing rate (ZCR), which represents the ratio where the sign of the data changes within a time interval. We propose a decision-making method based on an estimate of the ZCR to enhance the ability to predict the stock price range, and apply our method to the Standard & Poors 500 (S&P500) stock index. The results indicate that our method can achieve better outcomes than other methods.
{"title":"Stock Price Range Forecast via a Recurrent Neural Network Based on the Zero-Crossing Rate Approach","authors":"Yu-Fei Lin, Yeong-Luh Ueng, W. Chung, Tzu-Ming Huang","doi":"10.1109/CIFEr.2019.8759061","DOIUrl":"https://doi.org/10.1109/CIFEr.2019.8759061","url":null,"abstract":"By knowing the future price range, which is the difference between the closing price and the opening price, we can calculate the long or short positions in advance. This paper presents a Recurrent Neural Network (RNN) based approach to forecast the price range. Compared to other methods based on machine learning, our method puts greater focus on the characteristics of the stock data, such as the zero-crossing rate (ZCR), which represents the ratio where the sign of the data changes within a time interval. We propose a decision-making method based on an estimate of the ZCR to enhance the ability to predict the stock price range, and apply our method to the Standard & Poors 500 (S&P500) stock index. The results indicate that our method can achieve better outcomes than other methods.","PeriodicalId":368382,"journal":{"name":"2019 IEEE Conference on Computational Intelligence for Financial Engineering & Economics (CIFEr)","volume":"35 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124135297","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：