Predict Price Movements Research Articles

TECHNICAL ANALYSIS AND ITS THEORETICAL BASIS FOR TRADING ACTIVITY MANAGEMENT

Research background: In today's era of modern technology, traditional and new methods as well as processes and technologies increasingly contrast. In the first chapter, we focused on a detailed description of the first historical changes in the emergence of trade and financial derivatives. We also identified and described the emergence of specialised trading and stock exchange locations. We continued with a description of the basic methods of analysis and focused on the basics of technical analysis. In the next chapter, we evaluated the current state of research in this area and, based on previous research, we classified and structured individual studies. Purpose of the article: The aim of our contribution is to examine, in detail, the basics and current research in the field of technical analysis. Methods: We consider the inconsistent results of the studies analysed to be very interesting. We consider the unsystematic and inconsistent methodology of the assessment of individual studies to be a major problem. As we mentioned in the article, in many instances the studies analysed focused only on individual aspects of technical analysis and evaluated the final results of the business process. Findings & Value added: In other cases, the results were skewed, mainly due to the nonacceptance of transaction costs or due to risk abstraction. However, despite such results, technical analysis as such can be a valuable tool for predicting price movements. Subsequently, we described our findings, interpreted them, and identified their limitations. We consider the aim of the study to have been met, and believe that it will be beneficial in the field of research on this issue.

The Prediction of Gold Price Movement by Comparing Naive Bayes, Support Vector Machine, and K-NN

Gold is a yellow precious metal that can be forged so it is easy to form with various forms of jewelry such as pendants, earrings, rings, bracelets and others, gold has a high value. Gold itself is an exchange rate used in ancient times before the existence of money as it is today. Gold also can be used as an investment that is profitable for the investor and it has less risks. Investment is a form of fund management to give benefit by putting fund in allocation that is predicted will give additional benetifs. Prediction of gold price movements or predictions of gold price in gold stock investment, this research uses 3 (three) algorithms that will be implemented in analysis and increase accuracy, in the discussion or research that was made using the Naïve Bayes algorithm, Support Vector Machine and K-Nearest Neighbor, the dataset is obtained from the website, namely www.finance.yahoo.com the data was then tested using Rapid miner tools so that the average value of the Support Vector Machine algorithm with an accuracy rate of 57.59%, precision 58 ,73% and recall 51,78%. The next is the Naïve Bayes algorithm so that it is known to have an accuracy rate of 55.59%, precision 54.55% and recall 51.70%. Based on the comparison of the three algorithms, it is known that the one with the best accuracy, precision, and recall is the K-NN algorithm with 61.90% accuracy, 60.98% precision, and 60.35% recall. Furthermore, the results of testing the K-Nearst Neighbor algorithm have good results compared to the 3 (three) other algorithm tests and the Naïve Bayes algorithm testing has a low level of accuracy, namely 55.59%, precision 54.55% and recall 51.70%. The research uses 3 algorithms, namely naive bayes, K-nearst neighbor and Support Vector Machine, because the three algorithms are well-established algorithms to be applied to research, especially in time series gold price research and are very good, especially for classification

A New Hybrid VMD-ICSS-BiGRU Approach for Gold Futures Price Forecasting and Algorithmic Trading

The gold market plays a vital role in the world economy. Due to its complex and nonstationary nature, predicting the price of gold is particularly challenging. In this study, a new hybrid forecasting approach named variational mode decomposition (VMD)-iterated cumulative sums of squares (ICSS)-bidirectional gated recurrent unit (BiGRU) is proposed by integrating BiGRU deep learning model, VMD, and iterated cumulative sum of squares algorithm. The forecasting framework is able to extract the inner factors and patterns within the gold futures market movements, decompose its correlation with external markets and detect shifts within market conditions in order to accurately predict price movements in the gold futures market. The experimental results show that the hybrid forecasting approach can improve the prediction performance significantly in comparison to the benchmarks. Furthermore, we extend the proposed hybrid forecasting approach to generate trading strategies and test trading performance of the gold futures market. The testing results over an out-of-sample period of 11 years (2008–2019) indicate that the strategy generated based on the prediction of the proposed approach displays high levels of consistency in generating positive returns and outperforms several other common trading strategies under various market conditions. The approach also shows consistent better results when generalized to the spot gold market, providing practical guidance for minimizing investment risk and hedging strategies in the gold commodity market.

Stock Price Movement Prediction Using Sentiment Analysis and CandleStick Chart Representation.

Determining the price movement of stocks is a challenging problem to solve because of factors such as industry performance, economic variables, investor sentiment, company news, company performance, and social media sentiment. People can predict the price movement of stocks by applying machine learning algorithms on information contained in historical data, stock candlestick-chart data, and social-media data. However, it is hard to predict stock movement based on a single classifier. In this study, we proposed a multichannel collaborative network by incorporating candlestick-chart and social-media data for stock trend predictions. We first extracted the social media sentiment features using the Natural Language Toolkit and sentiment analysis data from Twitter. We then transformed the stock’s historical time series data into a candlestick chart to elucidate patterns in the stock’s movement. Finally, we integrated the stock’s sentiment features and its candlestick chart to predict the stock price movement over 4-, 6-, 8-, and 10-day time periods. Our collaborative network consisted of two branches: the first branch contained a one-dimensional convolutional neural network (CNN) performing sentiment classification. The second branch included a two-dimensional (2D) CNN performing image classifications based on 2D candlestick chart data. We evaluated our model for five high-demand stocks (Apple, Tesla, IBM, Amazon, and Google) and determined that our collaborative network achieved promising results and compared favorably against single-network models using either sentiment data or candlestick charts alone. The proposed method obtained the most favorable performance with 75.38% accuracy for Apple stock. We also found that the stock price prediction achieved more favorable performance over longer periods of time compared with shorter periods of time.

Generative adversarial network (GAN) and enhanced root mean square error (ERMSE): deep learning for stock price movement prediction

The prediction of stock price movement direction is significant in financial circles and academic. Stock price contains complex, incomplete, and fuzzy information which makes it an extremely difficult task to predict its development trend. Predicting and analysing financial data is a nonlinear, time-dependent problem. With rapid development in machine learning and deep learning, this task can be performed more effectively by a purposely designed network. This paper aims to improve prediction accuracy and minimizing forecasting error loss through deep learning architecture by using Generative Adversarial Networks. It was proposed a generic model consisting of Phase-space Reconstruction (PSR) method for reconstructing price series and Generative Adversarial Network (GAN) which is a combination of two neural networks which are Long Short-Term Memory (LSTM) as Generative model and Convolutional Neural Network (CNN) as Discriminative model for adversarial training to forecast the stock market. LSTM will generate new instances based on historical basic indicators information and then CNN will estimate whether the data is predicted by LSTM or is real. It was found that the Generative Adversarial Network (GAN) has performed well on the enhanced root mean square error to LSTM, as it was 4.35% more accurate in predicting the direction and reduced processing time and RMSE by 78 secs and 0.029, respectively. This study provides a better result in the accuracy of the stock index. It seems that the proposed system concentrates on minimizing the root mean square error and processing time and improving the direction prediction accuracy, and provides a better result in the accuracy of the stock index.

Forecasting price movements of global financial indexes using complex quantitative financial networks

As predicting trends in the financial market becomes more important, and artificial intelligence technology advances, there is active research on predicting stock movements by analyzing historical prices. However, few attempts have been made to utilize complex financial networks to predict price movements. Most studies focus only on the target financial index, and only a few studies examine both the target financial index and influential financial indexes. To fill the research gap, we propose a novel deep learning algorithm based on quantitative complex financial networks to forecast the price movement of global financial indexes using technical analysis. In other words, we propose a method that analyses the causal relationship between financial indexes in a cleaned correlation network, which considers the causal impact in this quantitatively constructed network. We use the random matrix theory to construct the financial network first, and we use transfer entropy to find directional impact within the network. Based on a daily historical dataset for global indexes and out-of-sample tests, the results show that the proposed method outperforms past state-of-the-art algorithms. Our findings reveal that identifying and using proper financial networks are important in predicting problems. Our study suggests that it is important to develop deep learning algorithms and to consider the financial network based on complex system theory when solving prediction problems in the financial market.

Stock movement prediction via gated recurrent unit network based on reinforcement learning with incorporated attention mechanisms

The recent advances usually mine market information from the chaotic data to conduct a stock movement prediction task. However, the current stock price movement prediction approaches mainly compute attention weighted sum of the global contextual semantic embeddings, which fails to combine local word-level or char-level ones to jointly learn news-level representation. Moreover, for Chinese stock price movement prediction task, some collected news texts are chaotic even irrelevant to the target stock. It suggests that the models need filter some news-level representations (viewed as noises) to enhance the performance. To that aim, we develop a novel stock price movement prediction network via bidirectional gated recurrent unit (GRU) network based on reinforcement learning (RL) with incorporated attention mechanism. In specific, to reduce the noise of news texts and learn news-level representation with more abundant semantics, two novel attention mechanisms respectively based on add and dot operation were first proposed in this work. We then design a novel GRU structure based on RL to filter some irrelated news-level representations (i.e., news-level noises) and capture abundant long-term dependencies. Finally, the experimental results show that the proposed model far outperforms the recent advances and achieves state-of-the-art performances.

Formula omitted]: COVID-19 adopted Hybrid and Parallel deep information fusion framework for stock price movement prediction

The novel of coronavirus (COVID-19) has suddenly and abruptly changed the world as we knew at the start of the 3rd decade of the 21st century. Particularly, COVID-19 pandemic has negatively affected financial econometrics and stock markets across the globe. Artificial Intelligence (AI) and Machine Learning (ML)-based prediction models, especially Deep Neural Network (DNN) architectures, have the potential to act as a key enabling factor to reduce the adverse effects of the COVID-19 pandemic and future possible ones on financial markets. In this regard, first, a unique COVID-19 related PRIce MOvement prediction (COVID19 PRIMO) dataset is introduced in this paper, which incorporates effects of social media trends related to COVID-19 on stock market price movements. Afterwards, a novel hybrid and parallel DNN-based framework is proposed that integrates different and diversified learning architectures. Referred to as the COVID-19 adopted Hybrid and Parallel deep fusion framework for Stock price Movement Prediction (COVID19-HPSMP), innovative fusion strategies are used to combine scattered social media news related to COVID-19 with historical mark data. The proposed COVID19-HPSMP consists of two parallel paths (hence hybrid), one based on Convolutional Neural Network (CNN) with Local/Global Attention modules, and one integrated CNN and Bi-directional Long Short term Memory (BLSTM) path. The two parallel paths are followed by a multilayer fusion layer acting as a fusion center that combines localized features. Performance evaluations are performed based on the introduced COVID19 PRIMO dataset illustrating superior performance of the proposed framework.

Determination of VaR on BBRI Stocks and BMRI Stocks Using the ARIMA-GARCH Model

Stocks are investment instruments that are much in demand by investors as a basis in financial storage. Return and risk are the most important things in investing. Return is a complete summary of investment and the return series is easier to handle than the price series. The movement of risk of loss is obtained from stock investments with profits. One way to calculate risk is value-at-risk. The movement of stocks is used to form a time series so that the calculation of risk can use time series. The purpose of this study was to find out the Value-at-Risk value of BBRI and BMRI stock using the ARIMA-GARCH model. The data used in this study was the daily closing price for 3 years. The time series method used is the Autoregressive Integrated Moving Average (ARIMA)-Generalized Autoregressive Conditional Heteroscedastic (GARCH) model. The stage of analysis is to determine the prediction of stock price movements using the ARIMA model used for the mean model and the GARCH model is used for volatility models. The average value and variants obtained from the model are used to calculate value-at-risk in BBRI and BMRI stock. The results obtained are the ARIMA(3,0,3)-GARCH(1,1) and ARIMA(2,0,2)-GARCH(1,1) model so with a significance level of 5% obtained Value-at-Risk of 0.04058 to BBRI stock and 0.10167 to BMRI stock.

House Price Predictability and Demographic Structure: Evidence from a Long-horizon International Data (인구구조와 주택가격 예측: 장기 패널데이터를 이용한 실증분석)

English Abstract: We examine the predictability of real house price movements along with demographic shifts, using a long time series international panel data set that covers 17 advanced countries over the period of 1950-2015. Since demographic structures change slowly and housing markets adjust slowly to economic fundamentals, we conduct our analysis over forecast horizons of 1, 2, 5, and 10 years. We find that (1) two most relevant demographic predictors for the future house price movements both over short- and long-horizon are the proportion of ages 70 and above in population of 20 years old and above and the proportion of ages 25-44; and (2) if samples are spanning over relatively short period in time, one can be misled about the house price predictability of demographic structures. These new results help to reconcile mixed evidence from the existing studies. Korean Abstract: 본 연구는 1950-2015년에 걸쳐 17 개 선진국을 포괄하는 장기 패널 자료를 사용하여 인구구조 변화가 미래 주택가격을 얼마나 예측하는지를 추정한다. 인구구조가비교적 천천히 변화하고 펀더멘털 변화에 대한 주택가격의 반응속도가 느릴 수 있는 점들을 고려하여 본 연구에서는 미래 주택가격 예측 기간을 1년, 2년, 5년, 10년으로 설정하고 각각 분석하였다. 본 연구의 주요 결과는 다음과 같다. 첫째, 단기와 장기 모두에서 미래 주택가격의 변동을 가장 잘 설명하는 두 개의 인구 변수는, 20세 이상 인구 중 70세 이상 인구가 차지하는 비율 및 25-44 세 인구가 차지하는 비율이다. 둘째, 이 같은 인구구조와 주택가격 사이의 관계는 기존 연구에서처럼 표본기간이 40년 전후로 비교적 짧은 경우 제대로 관찰되지 않는다. 이러한 결과들은 왜 기존 연구들이 서로 상충된 결과를 보고하거나 이론에 반하는 결과를 보고하는지를 이해하는데 도움을 준다.

A TOPSIS-ELM framework for stock index price movement prediction

In recent years Extreme Learning Machine (ELM) has gained the interest of various researchers due to its superior generalization and approximation capability. The network architecture and type of activation functions are the two important factors that influence the performance of an ELM. Hence in this study, a Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) oriented multi-criteria decision making (MCDM) framework is suggested for analyzing various ELM models developed with distinct activation functions with respect to sixteen evaluation criteria. Evaluating the performance of the ELM with respect to multiple criteria instead of single criterion can help in designing a more robust network. The proposed framework is used as a binary classification system for pursuing the problem of stock index price movement prediction. The model is empirically evaluated by using historical data of three stock indices such as BSE SENSEX, S&P 500 and NIFTY 50. The empirical study has disclosed promising results by evaluating ELM with different activation functions as well as multiple criteria.

Stock Market Prediction with Gaussian Naïve Bayes Machine Learning Algorithm

The stock market is one of the key sectors of a country’s economy. It provides investors with an opportunity to invest and gain returns on their investment. Predicting the stock market is a very challenging task and has attracted serious interest from researchers from many fields such as statistics, artificial intelligence, economics, and finance. An accurate prediction of the stock market reduces investment risk in the market. Different approaches have been used to predict the stock market. The performances of Machine learning (ML) models are typically superior to those of statistical and econometric models. The ability of Gaussian Naïve Bayes ML algorithm to predict stock price movement has not been addressed properly in the existing literature, hence this work attempt to fill that gap by evaluating the performance of GNB algorithm when combined with different feature scaling and feature extraction techniques in stock price movement prediction. The performance of the GNB models set up were ranked using the Kendall’s test of concordance for the various evaluation metrics used. The results indicated that, the predictive model based on integration of GNB algorithm and Linear Discriminant Analysis (GNB_LDA) outperformed all the other models of GNB considered in three of the four evaluation metrics (i.e., accuracy, F1-score, and AUC). Similarly, the predictive model based on integration of GNB algorithm, Min-Max scaling, and PCA produced the best rank using the specificity results. In addition, GNB produced better performance with Min-Max scaling technique than it does with standardization scaling techniques

Enhancing Pretrained Models with Domain Knowledge

Unsupervised pretraining has recently achieved significant success on a wide variety of natural language processing tasks. An important problem that remains understudied is how to effectively adapt such models to specific domains, which is critical for many real-life applications. In this paper, we explore to enhance pretraining by leveraging two typical sources of domain knowledge: unstructured domain-specific text and structured (often human-curated) domain knowledge. We propose models to jointly utilize these different sources of knowledge, which achieve the state-of-the-art results on two tasks of different domains: stock price movement prediction and software bug duplication detection, by adapting publicly available pretrained models obtained on generic domain-free corpora like book corpora and news articles.

Accuracy in Market Price and Real-Time Updates of Chart Through Computer Usage to the Inclusive Financial Market World

This paper looked at accuracy in market price and real-time updates of chart through computer usage to the inclusive financial market world. It also examined the revolution in the advancement of technology based on the application of computer in the financial market world. The evolution of computer devices has created an open opportunity for more and more participants in the financial market. The profitability of investments in financial markets highly depends on the predictability of price movements. If a forecasting model or technique can precisely predict the direction of the market, investment risk and uncertainty can be minimized. Market execution are instant which now produces immerse gain to the traders, buying, selling and holding currencies in order to make a profit from favorable fluctuations in exchange rates. Trading The forex market is one of the most popular markets for speculation, due to its enormous size, liquidity, and tendency for currencies to move in strong trends.

Stock Price Movement Prediction Based on a Deep Factorization Machine and the Attention Mechanism

The prediction of stock price movement is a popular area of research in academic and industrial fields due to the dynamic, highly sensitive, nonlinear and chaotic nature of stock prices. In this paper, we constructed a convolutional neural network model based on a deep factorization machine and attention mechanism (FA-CNN) to improve the prediction accuracy of stock price movement via enhanced feature learning. Unlike most previous studies, which focus only on the temporal features of financial time series data, our model also extracts intraday interactions among input features. Further, in data representation, we used the sub-industry index as supplementary information for the current state of the stock, since there exists stock price co-movement between individual stocks and their industry index. The experiments were carried on the individual stocks in three industries. The results showed that the additional inputs of (a) the intraday interactions among input features and (b) the sub-industry index information effectively improved the prediction accuracy. The highest prediction accuracy of the proposed FA-CNN model is 64.81%. It is 7.38% higher than that of traditional LSTM, and 3.71% higher than that of the model without sub-industry index as additional input features.

Bitcoin closing price movement prediction with optimal functional link neural networks

Bitcoin prediction is a recent area of research interest and growing fast. Since its inception, in a short period of time Bitcoin got wide popularity and considered as an investment asset. At present it is the most successful cryptocurrency compared to other altcoins dispersed in the world economy. The Bitcoin prices fluctuate like other stock markets due to inherent volatility. The investors’ confidence on Bitcoin rising fast and have been reflected on its prices. Though few computational intelligence methods are available, sophisticated methodologies for accurate prediction of Bitcoin are still lacking and need to be explored. Functional link neural network (FLN) is a flat network, offers lower computational complexity, and achieves enhanced input–output nonlinearity mapping through functional expansion of input signals. Basis functions such as Legendre, Trigonometric, Laguerre, and Chebyshev are commonly used polynomials in FLN for expansion of input signal dimension. In this article along with the weight and bias vector of FLNs, optimal number of polynomial functions for each category of basis function is selected by genetic algorithm during training process rather fixing them earlier. Therefore, an optimal FLN structure is crafted on fly from exploitation of training data. The optimal FLN models are used to predict the daily, weekly, and monthly closing prices of Bitcoin. A comparative study among optimal FLNs is carried out using evaluation metrics such as MAPE, NMSE, ARV, and U of Theil’s statistics. Finally, outcomes from experimental and comparative studies suggested the superiority of optimal FLN models for Bitcoin closing price prediction.

Comparative analysis of naïve bayes and knn on prediction of forex price movements for gbp/usd currency at time frame daily

This study aims to analyze the comparison of the Naïve Bayes and kNN on the Prediction of Forex Price Movements for GBP / USD on Time Frame Daily. The data used is taken from the metatrader-4 application which is often used by forex traders when making transactions. There are 2,145 data rows consisting of the date, hour, open price, high, low, close, and transaction volume columns. From this data, a column for the target class is created with the name ‘result’. The result column is filled with increasing or decreasing values. The value of increase or decrease is obtained from the comparison of the previous closing price with the closing price of the next day. This study analyzes the results of the comparison of the data mining classification of algorithm between the Naïve Bayes algorithm and kNN. The 2,145 data were divided into 2 parts, namely 80% for training data and 20% for testing data. The analysis is done by comparing the precision, recall, and accuracy test results for each algorithm. The conclusion of this study is that the kNN algorithm is better than the Naïve Bayes algorithm in case of predicting forex price movements for GBP/USD currency at time frame daily.

An end-to-end multitask method with two targets for high-frequency price movement prediction

An end-to-end multitask method with two targets for high-frequency price movement prediction

Cluster-Based Mutual Fund Classification and Price Prediction Using Machine Learning for Robo-Advisors.

The rise of FinTech has been meteoric in China. Investing in mutual funds through robo-advisor has become a new innovation in the wealth management industry. In recent years, machine learning, especially deep learning, has been widely used in the financial industry to solve financial problems. This paper aims to improve the accuracy and timeliness of fund classification through the use of machine learning algorithms, that is, Gaussian hybrid clustering algorithm. At the same time, a deep learning-based prediction model is implemented to predict the price movement of fund classes based on the classification results. Fund classification carried out using 3,625 Chinese mutual funds shows both accurate and efficient results. The cluster-based spatiotemporal ensemble deep learning module shows better prediction accuracy than baseline models with only access to limited data samples. The main contribution of this paper is to provide a new approach to fund classification and price movement prediction to support the decision-making of the next generation robo-advisor assisted by artificial intelligence.

Predicting market movement direction for bitcoin: A comparison of time series modeling methods

Many traders participate in activities known as "day-trading", trading Bitcoin against the dollar bill as the United States Dollar (USD) on very short timeframes to squeeze out profits from small market fluctuations. This paper aims to help traders decide how to best act by creating a model that can predict price movement's direction for the next 5-min time frame. Several machine-learning models have been tested for this Up/Down binary-classification problem. In this paper, we provide a comparison of the state-of-art strategies in predicting the movement direction for bitcoin, including Random Guessing and a Momentum-Based Strategy. The tested models include Autoregressive Integrated Moving Average (ARIMA), Prophet (by Facebook), Random Forest, Random Forest Lagged-Auto-Regression, and Multi-Layer Perceptron (MLP) Neural Networks. The MLP deep neural network has achieved the highest accuracy of 54% compared to other time-series prediction models. Also, in this paper, various data transformation and feature engineering have been applied in the comparison.