This study utilizes the XGBoost algorithm in the field of machine learning to conduct quantitative stock picking research for CSI 300 stocks. The article firstly outlines the importance and practical application background of quantitative stock selection, and then discusses in depth the basic principle of XGBoost algorithm and its application method in quantitative stock selection. By collecting historical data of CSI 300 stocks and after data preprocessing, this study constructs a multi-factor stock prediction model based on XGBoost and conducts relevant backtesting. Comparative experiments show that the XGBoost algorithm exhibits good effectiveness and demonstrates the unique advantages and characteristics of its stock selection strategy. The conclusion of the study shows that the XGBoost-based stock selection strategy has potential application value in the stock market and can provide investors with accurate and efficient stock selection reference.
{"title":"An Example of Machine Learning-Based Multifactor Dynamic Quantitative Stock Picking Models","authors":"Haocheng Sun","doi":"10.61173/7n8qt956","DOIUrl":"https://doi.org/10.61173/7n8qt956","url":null,"abstract":"This study utilizes the XGBoost algorithm in the field of machine learning to conduct quantitative stock picking research for CSI 300 stocks. The article firstly outlines the importance and practical application background of quantitative stock selection, and then discusses in depth the basic principle of XGBoost algorithm and its application method in quantitative stock selection. By collecting historical data of CSI 300 stocks and after data preprocessing, this study constructs a multi-factor stock prediction model based on XGBoost and conducts relevant backtesting. Comparative experiments show that the XGBoost algorithm exhibits good effectiveness and demonstrates the unique advantages and characteristics of its stock selection strategy. The conclusion of the study shows that the XGBoost-based stock selection strategy has potential application value in the stock market and can provide investors with accurate and efficient stock selection reference.","PeriodicalId":438278,"journal":{"name":"Science and Technology of Engineering, Chemistry and Environmental Protection","volume":"317 4","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141381395","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}
Sentiment analysis of film reviews has been a popular research topic, and previous researchers have investigated it on the IMDb dataset using a variety of machine learning models, however, the classification results are not satisfactory. Therefore this study aims to construct an effective sentiment analysis model and explore whether the Random Forest algorithm can be applied to the task of sentiment analysis on the IMDb dataset. In this study, after preprocessing the data, the Random Forest model was trained using a training set and evaluated using a test set to explore the accuracy and performance of the Random Forest model in film review sentiment analysis. The study also plotted word clouds to visualize the decision-making effect of the model. The Random Forest Model achieves an impressive 86% accuracy in sentiment analysis, while the word cloud plots provide a visually appealing depiction of its classification task. This indicates that the Random Forest model performs well in the film review sentiment analysis task with high accuracy and performance.
{"title":"Sentiment Analysis for Film Reviews Based on Random Forest","authors":"Dongling Zheng","doi":"10.61173/5t8epb44","DOIUrl":"https://doi.org/10.61173/5t8epb44","url":null,"abstract":"Sentiment analysis of film reviews has been a popular research topic, and previous researchers have investigated it on the IMDb dataset using a variety of machine learning models, however, the classification results are not satisfactory. Therefore this study aims to construct an effective sentiment analysis model and explore whether the Random Forest algorithm can be applied to the task of sentiment analysis on the IMDb dataset. In this study, after preprocessing the data, the Random Forest model was trained using a training set and evaluated using a test set to explore the accuracy and performance of the Random Forest model in film review sentiment analysis. The study also plotted word clouds to visualize the decision-making effect of the model. The Random Forest Model achieves an impressive 86% accuracy in sentiment analysis, while the word cloud plots provide a visually appealing depiction of its classification task. This indicates that the Random Forest model performs well in the film review sentiment analysis task with high accuracy and performance.","PeriodicalId":438278,"journal":{"name":"Science and Technology of Engineering, Chemistry and Environmental Protection","volume":"108 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141377093","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}
This study addresses the predictive challenges in China’s A-share market, characterized by high retail investor participation and significant policy impacts. We introduce a novel predictive model that leverages both machine learning algorithms and sentiment analysis to forecast market trends. The research utilizes comprehensive datasets, including real-time A-share market data and sentiment-derived data from stock-related news, processed via advanced machine learning techniques like Random Forest and sentiment analysis tools. Our approach innovatively combines traditional technical indicators with sentiment scores to enhance the predictive accuracy of the model. The findings suggest that integrating sentiment analysis significantly improves the model’s performance, evidenced by enhanced prediction metrics such as Mean Absolute Error (MAE) and R-squared values, which compare favorably before and after incorporating sentiment data. This study not only contributes to the existing financial prediction literature by providing a hybrid methodological approach but also offers practical implications for investors and policymakers in navigating the volatile A-share market.
本研究探讨了中国 A 股市场的预测难题，该市场的特点是散户投资者参与度高且受政策影响较大。我们引入了一个新颖的预测模型，利用机器学习算法和情感分析来预测市场趋势。研究利用综合数据集，包括实时 A 股市场数据和股票相关新闻的情绪衍生数据，并通过随机森林等先进的机器学习技术和情绪分析工具进行处理。我们的方法创新性地将传统技术指标与情感评分相结合，以提高模型的预测准确性。研究结果表明，整合情感分析可显著提高模型的性能，这体现在平均绝对误差（MAE）和 R 平方值等预测指标的增强上，在整合情感数据之前和之后，这些指标的对比结果都很好。这项研究提供了一种混合方法，不仅为现有的金融预测文献做出了贡献，还为投资者和政策制定者驾驭动荡的 A 股市场提供了实际意义。
{"title":"A-share Trend Prediction Based on Machine Learning and Sentiment Analysis","authors":"Jiaming Zhang","doi":"10.61173/52jy4c52","DOIUrl":"https://doi.org/10.61173/52jy4c52","url":null,"abstract":"This study addresses the predictive challenges in China’s A-share market, characterized by high retail investor participation and significant policy impacts. We introduce a novel predictive model that leverages both machine learning algorithms and sentiment analysis to forecast market trends. The research utilizes comprehensive datasets, including real-time A-share market data and sentiment-derived data from stock-related news, processed via advanced machine learning techniques like Random Forest and sentiment analysis tools. Our approach innovatively combines traditional technical indicators with sentiment scores to enhance the predictive accuracy of the model. The findings suggest that integrating sentiment analysis significantly improves the model’s performance, evidenced by enhanced prediction metrics such as Mean Absolute Error (MAE) and R-squared values, which compare favorably before and after incorporating sentiment data. This study not only contributes to the existing financial prediction literature by providing a hybrid methodological approach but also offers practical implications for investors and policymakers in navigating the volatile A-share market.","PeriodicalId":438278,"journal":{"name":"Science and Technology of Engineering, Chemistry and Environmental Protection","volume":"143 5‐6","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141381158","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}
This article investigates the obstacle avoidance technology of unmanned aerial vehicles (UAVs). Firstly, the background, purpose, and significance of UAV obstacle avoidance technology are introduced. Then, the domestic and foreign research status is analyzed. Next, an overall plan for UAV obstacle avoidance is proposed, including demand analysis and system design. Afterwards, the theory and specific steps of binocular stereo vision obstacle positioning are discussed in detail, including camera calibration, image rectification, and stereo matching. Furthermore, methods for estimating obstacle motion states are researched. In addition, path planning methods for obstacle avoidance are explored, with a focus on the principles, issues, and improvement methods of artificial potential field method. Finally, the main achievements of this study are summarized, and future research directions are outlined. The innovation of this article lies in the proposal of an improved artificial potential field method, and the precise obstacle positioning achieved through binocular stereo vision. Future research can further optimize the path planning methods for obstacle avoidance to enhance the effectiveness and reliability of UAV obstacle avoidance.
{"title":"UAV Obstacle Avoidance Technology","authors":"Hongyi Wang","doi":"10.61173/vpv1ca75","DOIUrl":"https://doi.org/10.61173/vpv1ca75","url":null,"abstract":"This article investigates the obstacle avoidance technology of unmanned aerial vehicles (UAVs). Firstly, the background, purpose, and significance of UAV obstacle avoidance technology are introduced. Then, the domestic and foreign research status is analyzed. Next, an overall plan for UAV obstacle avoidance is proposed, including demand analysis and system design. Afterwards, the theory and specific steps of binocular stereo vision obstacle positioning are discussed in detail, including camera calibration, image rectification, and stereo matching. Furthermore, methods for estimating obstacle motion states are researched. In addition, path planning methods for obstacle avoidance are explored, with a focus on the principles, issues, and improvement methods of artificial potential field method. Finally, the main achievements of this study are summarized, and future research directions are outlined. The innovation of this article lies in the proposal of an improved artificial potential field method, and the precise obstacle positioning achieved through binocular stereo vision. Future research can further optimize the path planning methods for obstacle avoidance to enhance the effectiveness and reliability of UAV obstacle avoidance.","PeriodicalId":438278,"journal":{"name":"Science and Technology of Engineering, Chemistry and Environmental Protection","volume":"102 9","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141377997","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}
Pneumonia is a serious disease that poses a threat to people’s health of all ages. It could happen when people are infected by viruses, fungi, bacteria etc. Typically, Chest X-rays are the first and foremost imaging approach to implement pneumonia detection. This paper introduces the latest research achievements to help those who are new in this field to have basic intuition about AI in pneumonia detection., including Vision Transformers on chest X-ray, a novel model based on RetinaNet, CP_DeepNet, and a novel Efficient NetV2L model. The last part gives some suggestions about the future study of pneumonia detection using Deep learning.
肺炎是一种严重的疾病，对所有年龄段的人的健康都构成威胁。当人们受到病毒、真菌、细菌等感染时，就有可能患上肺炎。通常情况下，胸部 X 光是检测肺炎的首要成像方法。本文介绍了最新的研究成果，以帮助初涉此领域的人员对人工智能在肺炎检测中的应用有基本的直观认识，包括胸部 X 射线上的视觉变换器、基于 RetinaNet 的新型模型、CP_DeepNet 和新型 Efficient NetV2L 模型。最后一部分对未来利用深度学习进行肺炎检测的研究提出了一些建议。
{"title":"A literature review of pneumonia detection algorithms based on deep learning and chest X-ray imaging","authors":"Chenyu Wang","doi":"10.61173/wpf17b07","DOIUrl":"https://doi.org/10.61173/wpf17b07","url":null,"abstract":"Pneumonia is a serious disease that poses a threat to people’s health of all ages. It could happen when people are infected by viruses, fungi, bacteria etc. Typically, Chest X-rays are the first and foremost imaging approach to implement pneumonia detection. This paper introduces the latest research achievements to help those who are new in this field to have basic intuition about AI in pneumonia detection., including Vision Transformers on chest X-ray, a novel model based on RetinaNet, CP_DeepNet, and a novel Efficient NetV2L model. The last part gives some suggestions about the future study of pneumonia detection using Deep learning.","PeriodicalId":438278,"journal":{"name":"Science and Technology of Engineering, Chemistry and Environmental Protection","volume":"30 51","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141378711","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}
The discovery of drugs is recognized as a lengthy, highly costly, and extremely complex process. For example, some traditional drug discovery methods consist of millions of trials to get a druggable compound to the market. Drug discovery based on artificial intelligence can be a prompt, low-cost, and effective way to streamline drug discovery. Although some works have been proposed to use artificial intelligence tools for drug discovery, few people summarize these advances in a systematic way. In this paper, we propose an organized and comprehensive review that outlines a broad range of appliances of artificial intelligence in drug discovery including harnessing virtual screening and molecular docking techniques, utilizing pathway networks for repurposing existing drugs, lead identification, biomarker research, identification of the target, diverse variety of artificial intelligence and their comparison, etc. In addition, we shed light on predicted limitations and challenges in drug discovery based on artificial intelligence, as well as sketch the strategies to harness its potential for upcoming drug design endeavors.
{"title":"A Review: Drug Discovery Methods Based on Artificial Intelligence","authors":"Jinhao Chen","doi":"10.61173/khh85h64","DOIUrl":"https://doi.org/10.61173/khh85h64","url":null,"abstract":"The discovery of drugs is recognized as a lengthy, highly costly, and extremely complex process. For example, some traditional drug discovery methods consist of millions of trials to get a druggable compound to the market. Drug discovery based on artificial intelligence can be a prompt, low-cost, and effective way to streamline drug discovery. Although some works have been proposed to use artificial intelligence tools for drug discovery, few people summarize these advances in a systematic way. In this paper, we propose an organized and comprehensive review that outlines a broad range of appliances of artificial intelligence in drug discovery including harnessing virtual screening and molecular docking techniques, utilizing pathway networks for repurposing existing drugs, lead identification, biomarker research, identification of the target, diverse variety of artificial intelligence and their comparison, etc. In addition, we shed light on predicted limitations and challenges in drug discovery based on artificial intelligence, as well as sketch the strategies to harness its potential for upcoming drug design endeavors.","PeriodicalId":438278,"journal":{"name":"Science and Technology of Engineering, Chemistry and Environmental Protection","volume":"9 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141376851","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}
The advancement of urbanization has brought urban land use change into the spotlight of research. Remote sensing technology, as the primary means of capturing this change, provides valuable data for research. The main research object of this paper is the four classification methods of land use in cities based on remote sensing technology, and it outlines the main roles and applications of these four classification methods. This paper concludes that the visual interpretation method is less efficient but widely used. In addition, the accuracy of supervised classification and deep learning methods for classification is low. Unsupervised classification is simple, but the results differ greatly from reality. The above commonly used land classification methods have their advantages and disadvantages, and their combined use ensures that remote sensing technology provides real-time, effective, and comprehensive land use information for urban planning and management. Finally, this paper looks forward to the combination of remote sensing technology and modern technology, such as artificial intelligence. The integrated technology can provide more accurate and efficient technical support for future urban development.
{"title":"Research on Urban Land Use based on Remote Sensing Technology","authors":"Xiaochun Xu","doi":"10.61173/9smr3z05","DOIUrl":"https://doi.org/10.61173/9smr3z05","url":null,"abstract":"The advancement of urbanization has brought urban land use change into the spotlight of research. Remote sensing technology, as the primary means of capturing this change, provides valuable data for research. The main research object of this paper is the four classification methods of land use in cities based on remote sensing technology, and it outlines the main roles and applications of these four classification methods. This paper concludes that the visual interpretation method is less efficient but widely used. In addition, the accuracy of supervised classification and deep learning methods for classification is low. Unsupervised classification is simple, but the results differ greatly from reality. The above commonly used land classification methods have their advantages and disadvantages, and their combined use ensures that remote sensing technology provides real-time, effective, and comprehensive land use information for urban planning and management. Finally, this paper looks forward to the combination of remote sensing technology and modern technology, such as artificial intelligence. The integrated technology can provide more accurate and efficient technical support for future urban development.","PeriodicalId":438278,"journal":{"name":"Science and Technology of Engineering, Chemistry and Environmental Protection","volume":"64 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141381565","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}
Through daily life, complex tasks require the neural encoding of spatial location for oneself and others. Previous research studies in rodents have shown that rodents have neural representations of themselves in addition to other rodents. ( Ovchinnikov, 2010) However, there is still a need to understand how the human brain processes spatial location for itself and others. Furthermore, it is important to research which parts of human cognition can affect location encoding mechanisms. The current study uses existing data to determine the correlation between environmental boundaries and human brain activity. Using spatial observation and navigation tasks, the study investigated whether a physical boundary can affect neural encoding using implanted electrodes, representing the participants’ location and others’ location while in a closed environment. Results showed that representations were strengthened when the encoding of location had a greater behavioral significance and was contingent upon the momentary cognitive state of the individual.Together, these findings support the existence of a shared encoding mechanism within the human brain that signifies the whereabouts of both individuals in communal settings. Moreover, they illuminate novel insights into the neural processes that govern spatial navigation and the perception of others in practical situations.
{"title":"Real world ambulatory boundary effect within MTL oscillation during movement in human brains","authors":"Siwei Wu","doi":"10.61173/62mqmr19","DOIUrl":"https://doi.org/10.61173/62mqmr19","url":null,"abstract":"Through daily life, complex tasks require the neural encoding of spatial location for oneself and others. Previous research studies in rodents have shown that rodents have neural representations of themselves in addition to other rodents. ( Ovchinnikov, 2010) However, there is still a need to understand how the human brain processes spatial location for itself and others. Furthermore, it is important to research which parts of human cognition can affect location encoding mechanisms. The current study uses existing data to determine the correlation between environmental boundaries and human brain activity. Using spatial observation and navigation tasks, the study investigated whether a physical boundary can affect neural encoding using implanted electrodes, representing the participants’ location and others’ location while in a closed environment. Results showed that representations were strengthened when the encoding of location had a greater behavioral significance and was contingent upon the momentary cognitive state of the individual.Together, these findings support the existence of a shared encoding mechanism within the human brain that signifies the whereabouts of both individuals in communal settings. Moreover, they illuminate novel insights into the neural processes that govern spatial navigation and the perception of others in practical situations.","PeriodicalId":438278,"journal":{"name":"Science and Technology of Engineering, Chemistry and Environmental Protection","volume":"14 13‐14","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141380154","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}
With the development of technology, in order to improve the user’s driving experience and driving safety, there are more and more vehicle tasks with high delay requirements. Therefore, lots of researchers have paid attention to task offloading scheduling.However, as vehicle tasks become increasingly complex, a single task may consist of multiple subtasks with dependencies between them.The complex data dependencies within them make it more and more difficult to design appropriate task offloading strategies. Considering that this problem is closely related to the scenarios and requirements in the real world, this study focuses on the design of task offloading decisions in the scenario of UAV-assisted vehicle network, in which MEC servers are installed in the macro base station and UAV to provide computing resources for vehicles. We designed a task offloading strategy based on MATD3 algorithm to deal with this problem. Following simulation trials, it is evident that our approach offers notable benefits in terms of both delay and energy usage.
{"title":"Research on UAV-assisted Vehicle networking task unloading strategy based on multi-agent reinforcement learning","authors":"Fanjin Zeng","doi":"10.61173/ceadt415","DOIUrl":"https://doi.org/10.61173/ceadt415","url":null,"abstract":"With the development of technology, in order to improve the user’s driving experience and driving safety, there are more and more vehicle tasks with high delay requirements. Therefore, lots of researchers have paid attention to task offloading scheduling.However, as vehicle tasks become increasingly complex, a single task may consist of multiple subtasks with dependencies between them.The complex data dependencies within them make it more and more difficult to design appropriate task offloading strategies. Considering that this problem is closely related to the scenarios and requirements in the real world, this study focuses on the design of task offloading decisions in the scenario of UAV-assisted vehicle network, in which MEC servers are installed in the macro base station and UAV to provide computing resources for vehicles. We designed a task offloading strategy based on MATD3 algorithm to deal with this problem. Following simulation trials, it is evident that our approach offers notable benefits in terms of both delay and energy usage.","PeriodicalId":438278,"journal":{"name":"Science and Technology of Engineering, Chemistry and Environmental Protection","volume":"30 14","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141378579","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}
Securities trading has always been a high-risk, high-return domain. Investors seek high returns while endeavoring to minimize risks as much as possible. Therefore, stock price prediction has become a popular and immensely valuable research topic. This paper will use the ARMA model to forecast stock prices. Firstly, an analysis was conducted on selected stock, determining that the price sequence exhibits no seasonal effects but does display volatility effects. The trend is essentially linear, and the relationship between volatility effects and trends fits an additive model. Based on this, preprocessing was conducted by taking the three-day moving average sequence of the series to eliminate the volatility effects, yielding a clean sequence trend. Then, the trend was differenced once to obtain a stationary sequence. Subsequently, the appropriate ARIMA model order was determined by the (partial) autocorrelation plot of this stationary sequence, and the model was fitted to the stock for prediction, yielding satisfactory results. This indicates that the model can accurately forecast long-term trends, but the filtering of volatility effects prevents the prediction results from sensitively reflecting short-term fluctuations.
证券交易一直是一个高风险、高回报的领域。投资者在追求高回报的同时，也在尽可能地降低风险。因此，股票价格预测已成为一个热门且极具价值的研究课题。本文将使用 ARMA 模型来预测股票价格。首先，对选定的股票进行分析，确定价格序列没有季节效应，但有波动效应。趋势基本上是线性的，波动效应与趋势之间的关系符合加法模型。在此基础上，通过对序列的三天移动平均序列进行预处理，以消除波动效应，从而得到清晰的序列趋势。然后，对趋势进行一次差分，以获得静态序列。随后，根据该静态序列的（部分）自相关图确定适当的 ARIMA 模型阶数，并将该模型拟合到股票上进行预测，结果令人满意。这表明该模型可以准确预测长期趋势，但由于过滤了波动效应，预测结果无法灵敏反映短期波动。
{"title":"ARIMA Model-Based Research on Stock Price Prediction","authors":"Dong Liang","doi":"10.61173/nq5kv133","DOIUrl":"https://doi.org/10.61173/nq5kv133","url":null,"abstract":"Securities trading has always been a high-risk, high-return domain. Investors seek high returns while endeavoring to minimize risks as much as possible. Therefore, stock price prediction has become a popular and immensely valuable research topic. This paper will use the ARMA model to forecast stock prices. Firstly, an analysis was conducted on selected stock, determining that the price sequence exhibits no seasonal effects but does display volatility effects. The trend is essentially linear, and the relationship between volatility effects and trends fits an additive model. Based on this, preprocessing was conducted by taking the three-day moving average sequence of the series to eliminate the volatility effects, yielding a clean sequence trend. Then, the trend was differenced once to obtain a stationary sequence. Subsequently, the appropriate ARIMA model order was determined by the (partial) autocorrelation plot of this stationary sequence, and the model was fitted to the stock for prediction, yielding satisfactory results. This indicates that the model can accurately forecast long-term trends, but the filtering of volatility effects prevents the prediction results from sensitively reflecting short-term fluctuations.","PeriodicalId":438278,"journal":{"name":"Science and Technology of Engineering, Chemistry and Environmental Protection","volume":"21 10‐11","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-06-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141379855","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：