Análisis técnico de mercados de criptomoneda con redes neuronales convolucionales.
| dc.contributor.advisor | Pineda Ríos, Wilmer Darío | |
| dc.contributor.author | Caicedo Rueda, Andres Felipe | |
| dc.contributor.corporatename | Universidad Santo Tomás | spa |
| dc.date.accessioned | 2022-07-25T15:58:24Z | |
| dc.date.available | 2022-07-25T15:58:24Z | |
| dc.date.issued | 2022-07-13 | |
| dc.description | Con el análisis técnico de los mercados financieros se estudia la interacción del mercado usando principalmente gráficos para intentar pronosticar la tendencia del precio en el futuro. Sin embargo, aunque los instantes de interacción del mercado junto con sus indicadores técnicos se representan como imágenes, los esfuerzos de automatización de las estrategias de compra y venta de activos se han dirigido especialmente hacia los métodos algorítmicos. La naturaleza visual del análisis técnico invita a buscar alternativas de solución dentro del ámbito de la visión por computador usando técnicas de clasificación de imágenes y detección de objetos. En este trabajo se construyó una base de datos de 240 imágenes a partir de la información cronológica de las variables de interacción del mercado (Precio, Volumen) y algunos de los indicadores técnicos más populares (EMA, Bollinger, MACD, RSI). En dicha base se etiquetaron los objetos correspondientes a las clases de los eventos que conforman una estrategia de compra y venta (Cruce de EMA, Cruce del MACD). Posteriormente, con las imágenes etiquetadas de la base se entrenó un modelo de aprendizaje profundo de redes neuronales convolucionales (CNN) para detectar objetos de esas mismas clases en imágenes nunca vistas y así identificar señales que permitan tomar decisiones de compra y venta en el mercado de cripto activos con monedas como Bitcoin, Ethereum o ADA entre muchas otras. Como resultado del estudio se obtuvieron dos modelos para la detección de cruces de EMA y MACD, el primero de tipo EfficientDet0 con AP50 del 76.69% y el segundo de tipo EfficientDet4 con AP50 del 87.26%. Esta técnica de detección de objetos para diseñar e implementar estrategias de compra y venta es fácil e intuitiva y gracias a su nivel de abstracción podría ser usada con diferentes periodicidades, activos y mercados. | spa |
| dc.description.abstract | With the technical analysis of the financial markets, the interaction of the market is studied using mainly graphs to try to forecast the price trend in the future. However, although the moments of market interaction along with its technical indicators are represented as images, efforts to automate asset buying and selling strategies have been directed especially towards algorithmic methods. The visual nature of technical analysis invites us to search for alternative solutions within the field of computer vision using image classification and object detection techniques. In this work, a database of 240 images was built from the chronological information of the market interaction variables (Price, Volume) and some of the most popular technical indicators (EMA, Bollinger, MACD, RSI). In this database, the objects corresponding to the classes of the events that make up a buying and selling strategy (EMA crossover, MACD crossover) were labeled. Subsequently, with the labeled images of the base, a deep learning model of convolutional neural networks (CNN) was trained to detect objects of these same classes in images never seen before and thus identify signals that allow buying and selling decisions in the market of crypto assets with currencies such as Bitcoin, Ethereum or ADA among many others. As a result of the study, two models were obtained for the detection of EMA and MACD crossovers, the first of the EfficientDet0 type with AP50 of 76.69% and the second of the EfficientDet4 type with AP50 of 87.26%. This object detection technique to design and implement buying and selling strategies is easy and intuitive and thanks to its level of abstraction it could be used with different periodicities, assets and markets. | spa |
| dc.description.degreelevel | Maestría | spa |
| dc.description.degreename | Magister en Estadística Aplicada | spa |
| dc.format.mimetype | application/pdf | |
| dc.identifier.citation | Caicedo Rueda, A. F. (2022). Análisis técnico de mercados de criptomoneda con redes neuronales convolucionales. [Trabajo de maestría, Universidad Santo Tomás]. Repositorio institucional. | spa |
| dc.identifier.instname | instname:Universidad Santo Tomás | spa |
| dc.identifier.reponame | reponame:Repositorio Institucional Universidad Santo Tomás | spa |
| dc.identifier.repourl | repourl:https://repository.usta.edu.co | spa |
| dc.identifier.uri | http://hdl.handle.net/11634/46046 | |
| dc.language.iso | spa | |
| dc.publisher | Universidad Santo Tomás | spa |
| dc.publisher.branch | CRAI-USTA Bogotá | spa |
| dc.publisher.faculty | Facultad de Estadística | spa |
| dc.publisher.program | Maestría Estadística Aplicada | spa |
| dc.relation.references | Aggarwal, Charu C. (2018). Neural Networks and Deep Learning. A Textbook. Cham: Springer, pág. 497. ISBN: 978-3-319-94462-3. DOI: 10.1007/978-3-319-94463-0. | spa |
| dc.relation.references | Azhikodan, Akhil, Anvitha Bhat y Mamatha Jadhav (mayo de 2019). ((Stock Trading Bot Using Deep Reinforcement Learning)). En: Lecture Notes in Networks and Systems, págs. 41-49. DOI: https://doi.org/10.1007/978-981-10-8201-6_5. | spa |
| dc.relation.references | Birogul, Serdar, Günay Temür y Utku Köse (mayo de 2020). ((YOLO Object Recognition Algorithm and "Buy-Sell Decision" Model over 2D Candlestick Charts)). En: IEEE Access PP, págs. 1-1. DOI: 10.1109/ACCESS.2020.2994282. | spa |
| dc.relation.references | Brock, William, Josef Lakonishok y Blake LeBaron (1992). ((Simple Technical Trading Rules and the Stochastic Properties of Stock Returns)). En: The Journal of Finance 47.5, págs. 1731-1764. ISSN: 00221082, 15406261. URL: http://www.jstor.org/stable/2328994. | spa |
| dc.relation.references | Caicedo, Andres (2022). Python Technical Analysis Indicators. [Online; accessed 28-April-2022]. URL: https://github.com/a3caicedo/PythonTechnicalAnalysisIndicators. | spa |
| dc.relation.references | Chan, Louis K., Narasimhan Jegadeesh y Josef Lakonishok (1996). ((Momentum Strategies)). En: The Journal of Finance 51.5, págs. 1681-1713. DOI: https://doi.org/10.1111/j.1540- 6261.1996. | spa |
| dc.relation.references | Cohen, Naftali, Tucker Balch y Manuela Veloso (jul. de 2019). ((Trading via Image Classification)). En: CoRR abs/1907.10046. arXiv: 1907.10046. URL: http://arxiv.org/abs/1907.10046 | spa |
| dc.relation.references | Duan, Kaiwen y col. (2019). ((CenterNet: Keypoint Triplets for Object Detection)). En: CoRR ab-s/1904.08189. arXiv: 1904.08189. URL: http://arxiv.org/abs/1904.08189. | spa |
| dc.relation.references | Ertel, Wolfgang (2017). Introduction to Artificial Intelligence. 2.a ed. Springer. ISBN: 978-3-319-58486-7. | spa |
| dc.relation.references | Fawaz, Hassan Ismail y col. (jul. de 2019). ((Deep learning for time series classification: a review)). En: Data Mining and Knowledge Discovery 33. DOI: 10.1007/s10618-019-00619-1. | spa |
| dc.relation.references | Girshick, Ross (abr. de 2015). ((Fast r-cnn)). En: DOI: 10.1109/ICCV.2015.169. | spa |
| dc.relation.references | Girshick, Ross B. y col. (2013). ((Rich feature hierarchies for accurate object detection and semantic segmentation)). En: CoRR abs/1311.2524. arXiv: 1311.2524. URL: http://arxiv.org/abs/1311. 2524. | spa |
| dc.relation.references | He, Kaiming y col. (2016). ((Deep Residual Learning for Image Recognition)). En: 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), págs. 770-778. DOI: 10.1109/CVPR.2016.90. | spa |
| dc.relation.references | Jegadeesh, Narasimhan y Sheridan Titman (2001). ((Profitability of Momentum Strategies: An Evaluation of Alternative Explanations)). En: The Journal of Finance 56.2, págs. 699-720. ISSN: 00221082, 15406261. URL: http://www.jstor.org/stable/222579. | spa |
| dc.relation.references | Karim, Fazle y col. (2019). ((Multivariate LSTM-FCNs for time series classification)). En: Neural Networks 116, págs. 237-245. ISSN: 0893-6080. DOI: https://doi.org/10.1016/j.neunet.2019.04.014. URL: https://www.sciencedirect.com/science/article/pii/S0893608019301200. | spa |
| dc.relation.references | Krizhevsky, Alex, Ilya Sutskever y Geoffrey E. Hinton (2012). ((ImageNet Classification with Deep Convolutional Neural Networks)). En: Proceedings of the 25th International Conference on Neural Information Processing Systems - Volume 1. NIPS’12. Lake Tahoe, Nevada: Curran Associates Inc., págs. 1097-1105. | spa |
| dc.relation.references | Lai, Guokun y col. (2017). ((Modeling Long- and Short-Term Temporal Patterns with Deep Neural Networks)). En: CoRR abs/1703.07015. arXiv: 1703. 07015. URL: http: // arxiv. org/ abs/1703. 07015. | spa |
| dc.relation.references | Law, Hei y Jia Deng (2018). ((CornerNet: Detecting Objects as Paired Keypoints)). En: CoRR abs/1808.01244. arXiv: 1808.01244. URL: http://arxiv.org/abs/1808.01244. | spa |
| dc.relation.references | Lecun, Y. y col. (1998). ((Gradient-based learning applied to document recognition)). En: Proceedings of the IEEE 86.11, págs. 2278-2324. DOI: 10.1109/5.726791. | spa |
| dc.relation.references | Lin, Tsung-Yi y col. (2014). ((Microsoft COCO: Common Objects in Context)). En: CoRR abs/1405.0312. arXiv: 1405.0312. URL: http://arxiv.org/abs/1405.0312. | spa |
| dc.relation.references | Liu, Wei y col. (2015). ((SSD: Single Shot MultiBox Detector)). En: CoRR abs/1512.02325. arXiv: 1512. 02325. URL: http://arxiv.org/abs/1512.02325. | spa |
| dc.relation.references | Moskowitz, Tobias J., Yao Hua Ooi y Lasse Heje Pedersen (2012). ((Time series momentum)). En: Journal of Financial Economics 104.2. Special Issue on Investor Sentiment, págs. 228-250. ISSN: 0304-405X. DOI: https://doi.org/10.1016/j.jfineco.2011.11.003. URL: https://www.sciencedirect.com/science/article/pii/S0304405X11002613. | spa |
| dc.relation.references | Redmon, Joseph y col. (2015). ((You Only Look Once: Unified, Real-Time Object Detection)). En: CoRR abs/1506.02640. arXiv: 1506.02640. URL: http://arxiv.org/abs/1506.02640. | spa |
| dc.relation.references | Ren, Shaoqing y col. (2015). ((Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks)). En: CoRR abs/1506.01497. arXiv: 1506. 01497. URL: http: // arxiv. org/ abs/1506. 01497. | spa |
| dc.relation.references | Sim, Hyun, Hae Kim y Jae Ahn (feb. de 2019). ((Is Deep Learning for Image Recognition Applicable to Stock Market Prediction?)) En: Complexity 2019, págs. 1-10. DOI: 10.1155/2019/4324878. | spa |
| dc.relation.references | Simonyan, Karen y Andrew Zisserman (2015). ((Very Deep Convolutional Networks for Large-Scale Image Recognition)). En: 3rd International Conference on Learning Representations, ICLR 2015, San Diego, CA, USA, May 7-9, 2015, Conference Track Proceedings. Ed. por Yoshua Bengio y Yann LeCun. URL: http://arxiv.org/abs/1409.1556. | spa |
| dc.relation.references | Szegedy, Christian y col. (2015). ((Going deeper with convolutions)). En: 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), págs. 1-9. DOI: 10.1109/CVPR.2015.7298594. | spa |
| dc.relation.references | Tan, Mingxing y Quoc V. Le (2019). ((EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks)). En: CoRR abs/1905.11946. arXiv: 1905. 11946. URL: http: // arxiv. org/ abs/1905. 11946. | spa |
| dc.relation.references | Tan, Mingxing, Ruoming Pang y Quoc V. Le (2019). ((EfficientDet: Scalable and Efficient Object Detection)). En: CoRR abs/1911.09070. arXiv: 1911.09070. URL: http://arxiv.org/abs/1911.09070. | spa |
| dc.relation.references | Tensorflow.org (2022). Object Detection with TensorFlow Lite Model Maker. [Online; accessed 07-March-2022]. URL: https://www.tensorflow.org/lite/tutorials/model_maker_object_detection. | spa |
| dc.relation.references | Velay, Marc y Fabrice Daniel (2018). ((Stock Chart Pattern recognition with Deep Learning)). En: CoRR abs/1808.00418. arXiv: 1808.00418. URL: http://arxiv.org/abs/1808.00418. | spa |
| dc.relation.references | Wang, Zhiguang, Weizhong Yan y Tim Oates (2016). ((Time Series Classification from Scratch with Deep Neural Networks: A Strong Baseline)). En: CoRR abs/1611.06455. arXiv: 1611 . 06455. URL: http://arxiv.org/abs/1611.06455. | spa |
| dc.relation.references | Wu, Yangru, Ronald Balvers y Erik Gilliland (feb. de 2000). ((Mean Reversion across National Stock Markets and Parametric Contrarian Investment Strategies)). En: Journal of Finance 55, págs. 745-772. DOI: 10.1111/0022-1082.00225. | spa |
| dc.relation.references | Yildirim, Deniz Can, Ismail Hakkı Toroslu y Ugo Fiore (2021). ((Forecasting directional movement of Forex data using LSTM with technical and macroeconomic indicators)). En: Financial Innovation 7. | spa |
| dc.relation.references | Zeiler, Matthew D. y Rob Fergus (2014). ((Visualizing and Understanding Convolutional Networks)). En: Computer Vision – ECCV 2014. Ed. por David Fleet y col. Cham: Springer International Publishing, págs. 818-833. ISBN: 978-3-319-10590-1. | spa |
| dc.rights | Atribución-NoComercial-SinDerivadas 2.5 Colombia | |
| dc.rights.accessrights | info:eu-repo/semantics/openAccess | |
| dc.rights.coar | http://purl.org/coar/access_right/c_abf2 | |
| dc.rights.local | Abierto (Texto Completo) | spa |
| dc.rights.uri | http://creativecommons.org/licenses/by-nc-nd/2.5/co/ | |
| dc.subject.keyword | technical analysis | spa |
| dc.subject.keyword | convolutional neural networks | spa |
| dc.subject.keyword | object detection | spa |
| dc.subject.keyword | cryptocurrencies | spa |
| dc.subject.lemb | Estadística | spa |
| dc.subject.lemb | Mercados | spa |
| dc.subject.lemb | Transferencia electrónica de fondos | spa |
| dc.subject.proposal | análisis técnico | spa |
| dc.subject.proposal | redes neuronales convolucionales | spa |
| dc.subject.proposal | detección de objetos | spa |
| dc.subject.proposal | criptomonedas | spa |
| dc.title | Análisis técnico de mercados de criptomoneda con redes neuronales convolucionales. | spa |
| dc.type | master thesis | |
| dc.type.coar | http://purl.org/coar/resource_type/c_bdcc | |
| dc.type.coarversion | http://purl.org/coar/version/c_ab4af688f83e57aa | |
| dc.type.drive | info:eu-repo/semantics/masterThesis | |
| dc.type.local | Tesis de maestría | spa |
| dc.type.version | info:eu-repo/semantics/acceptedVersion |
Archivos
Bloque original
1 - 3 de 3
Cargando...
- Nombre:
- 2022andrescaicedo.pdf
- Tamaño:
- 3 MB
- Formato:
- Adobe Portable Document Format
- Descripción:
- Trabajo de Grado
Cargando...
- Nombre:
- CartaAprobacionFacultad.pdf
- Tamaño:
- 198.18 KB
- Formato:
- Adobe Portable Document Format
- Descripción:
- Carta aprobación facultad
Cargando...
- Nombre:
- CartaDerechosAutor.pdf
- Tamaño:
- 294.61 KB
- Formato:
- Adobe Portable Document Format
- Descripción:
- Carta derechos de autor
Bloque de licencias
1 - 1 de 1
Cargando...
- Nombre:
- license.txt
- Tamaño:
- 807 B
- Formato:
- Item-specific license agreed upon to submission
- Descripción:

