Leonardo Kanashiro Felizardo, Francisco Caio Lima Paiva, Catharine de Vita Graves, Élia Yathie Matsumoto · 7 authors
The interdisciplinary relationship between machine learning and financial markets has long been a theme of great interest among both research communities. Recently, reinforcement learning and deep learning methods gained prominence in the active asset trading task, aiming to achieve outstanding performances compared with classical benchmarks, such as the Buy and Hold strategy. This paper explores both the supervised learning and reinforcement learning approaches applied to active asset trading, drawing attention to the benefits of both approaches. This work extends the comparison between the supervised approach and reinforcement learning by using state-of-the-art strategies with both techniques. We propose adopting the ResNet architecture, one of the best deep learning approaches for time series classification, into the ResNet-LSTM actor (RSLSTM-A). We compare RSLSTM-A against classical and recent reinforcement learning techniques, such as recurrent reinforcement learning, deep Q-network, and advantage actor–critic. We simulated a currency exchange market environment with the price time series of the Bitcoin, Litecoin, Ethereum, Monero, Nxt, and Dash cryptocurrencies to run our tests. We show that our approach achieves better overall performance, confirming that supervised learning can outperform reinforcement learning for trading. We also present a graphic representation of the features extracted from the ResNet neural network to identify which type of characteristics each residual block generates.
Following the recent trend of data-centric AI, we propose a clustering method to offer additional insight into precious and yet less-explored cryptocurrency price time series. While invaluable efforts have been conveyed in the domain of time series clustering, we integrate and harmonize some of the best practices in the field, namely Gramian Angular Field (GAF), Variational AutoEncoders (VAEs), and Deep Embedded Clustering (DEC). We use time series to image transformations as a preprocessing step for VAE to reduce dimensionality. After performing K-means clustering on VAE’s latent space, we provide DEC with cluster centroids from the previous step and retrain our network to do the clustering task. We evaluate the proposed method with the Bitcoin Tick-bar price dataset from 2017 onwards. Results demonstrate that our method leads to financially interpretable clusters and can improve Silhouette Score up to 10 percent compared to non-imaged time series.
Since the appearance of Bitcoin, cryptocurrencies have experienced enormousgrowth not only in terms of capitalization but also in number. As a result, thecryptocurrency market can be an attractive arena for investors as it offers manypossibilities, but a difficult one to understand as well. In this work, we aim tosummarize and segment the whole cryptocurrency market in 2018 with the helpof data analysis tools. We will use three different partitional clustering algorithmseach of them using a different representation for cryptocurrencies, namely: yearlymean and standard deviation of the returns, distribution of returns, and timeseries of returns. Since each representation will provide a different andcomplementary perspective of the market, we will also explore the combination ofthe three clustering results to obtain a fine-grained analysis of the main trends ofthe market. Finally, we will analyse the association of the clustering results withother descriptive features of the cryptocurrencies, including the age, technologicalattributes, and financial ratios derived from them. This will help to enhance theprofiling of the clusters with additional insights. As a result, this work offers adescription of the market and a methodology that can be reproduced by investorsthat want to understand the main trends on the market and that look forcryptocurrencies with different financial performance.
Feb 23, 2021·Proceedings of the ... annual Precise Time and Time Interval (PTTI) Systems and Applications Meeting/Proceedings of a meeting - Precise Time and Time Interval (PTTI) Systems and Applications Meeting
"Transcript of Video Presentation" Legal time and traceability to official time sources are becoming increasingly important. Engineering for civil timekeeping purposes demands timestamps with sufficient metadata to accurately represent local time. Spatial-temporal geographic information systems (GIS), or 4D GIS, provide three dimensions of geographic coordinates to identify a physical space and the fourth dimension referring to time. Geographic location information significantly enhances the provenance of local timestamps. “Traceability” has different meanings. One is of markers for abstract elements such as data, intellectual properties, or events stating when and where those were recorded or taken place. The other is of evidence attached to physical assets attesting to their origin, their transit between locations, and chain of custody. These applications of traceability share universal ground-truths of earth-bound locations and UTC timekeeping. This paper describes an aggregate of two efforts that were presented at past PTTI conferences in 2019 and 2020: Common Calendar Timestamp system (CCT) by EdlMax and modern timekeeping for IT solutions by Sync-n-Scale. This collaboration introduces “geostamping” as a tool for meeting traceability needs across multitude of industry segments. A geostamp is an atomic data element consisting of a geographic location (e.g. cartesian latitude, longitude, and altitude) and CCT timestamp. Its authenticity is established by combining a demonstrably trusted geographic coordinate and a UTC-traceable accurate timestamp. A geostamp is human readable and technically accurate for interpretive and forensic purposes. It can be examined afterwards and away from its issuing system for determination of where and when it was recorded. This characteristic makes it suitable for general and legal purposes where time recording is used for tracking and auditing. Trusted geostamps are also suitable for a wide range of 4D GIS applications in machine learning, artificial intelligence, data analytics and Blockchain distributed ledgers. The CCT specification has been extended to include geographic coordinates to support geostamp. Organization of its code base has been refined to provide clear guidance on the interpretation of IERS Leap Seconds, IANA Time Zone Database and geographic information, algorithms for UTC calculation, binary and character timestamp formats, methods for populating and reading the two formats, and a high-level API for convenient manipulation and calculation of calendaring operations. The C/C++ reference implementation has been updated to accommodate the full breadth of civil time data provided by Tz database. It is now also available to the community on GitHub. The authors hope these improvements will find support in general adoption and look forward to engaging with the community for future features.
Traditional time series modeling techniques emphasize on predicting cryptocurrencies using classically structured data representation as numerical features to present the time-series datasets. In this paper, a novel approach to analyze time-series data charts using a modified Convolutional Neural Networks (CNNs) is proposed. The CNNs have been adopted to recognize subtle and undetectable patterns within images of time-series data charts. Our approach has been proven to achieve significant results, suggesting a need for further research into this new method for time series modeling, especially for Bitcoin.
Petrônio Cândido de Lima e Silva, Carlos Alberto Severiano, Marcos Antônio Alves, Rodrigo Silva · 6 authors
In this paper we introduce a Non-Stationary Fuzzy Time Series (NSFTS) method with time varying parameters adapted from the distribution of the data. In this approach, we employ Non-Stationary Fuzzy Sets, in which perturbation functions are used to adapt the membership function parameters in the knowledge base in response to statistical changes in the time series. The proposed method is capable of dynamically adapting its fuzzy sets to reflect the changes in the stochastic process based on the residual errors, without the need to retraining the model. This method can handle non-stationary and heteroskedastic data as well as scenarios with concept-drift. The proposed approach allows the model to be trained only once and remain useful long after while keeping reasonable accuracy. The flexibility of the method by means of computational experiments was tested with eight synthetic non-stationary time series data with several kinds of concept drifts, four real market indices (Dow Jones, NASDAQ, SP500 and TAIEX), three real FOREX pairs (EUR-USD, EUR-GBP, GBP-USD), and two real cryptocoins exchange rates (Bitcoin-USD and Ethereum-USD). As competitor models the Time Variant fuzzy time series and the Incremental Ensemble were used, these are two of the major approaches for handling non-stationary data sets. Non-parametric tests are employed to check the significance of the results. The proposed method shows resilience to concept drift, by adapting parameters of the model, while preserving the symbolic structure of the knowledge base.
Businesses whose work hours, work durations & also more compliance records must be trustworthy and verifiable have great difficulties maintaining more accurate & also safe timekeeping. This article looks at a novel approach for smart contract & also blockchain technology integration protection of timekeeping data. Changing from more conventional, centralized time-tracking systems to a distributed ledger creates a safe framework that logs time inputs as unchangeable transactions, therefore preventing any retroactive changes or more unauthorized additions. While smart contracts provide automation of activities like clock-ins, clock-outs & more compliance alerts, every input is validated & also documented with cryptographic certainty. This guarantees real-time policy compliance and eliminates errors in hand-made monitoring. Clear audit trail, real-time notifications for errors or transgressions, and tamper-evident recordkeeping define the main characteristics of the recommended solution. These elements taken together provide a mechanism wherein auditors, managers, and employees may trust the data free from middlemen's intervention. Extended data integrity, increased organizational trust, greater transparency in payroll and compliance procedures, and fewer incidence of fraud or human error define the projected benefits. From manufacturing to logistics, remote work, healthcare, legal compliance management, the approach is versatile and scalable across numerous sectors. This work integrates trust into the infrastructure via blockchain and smart contract automation therefore establishing a progressive base for future timekeeping systems
Psychology tests are widely used in mental health diagnoses, education assessments, and recruitment assessments. There are several major problems in using the traditional way to manage psychology test data. First, there is no single source of truth for psychology test data due to centralized storage. Second, the data are mutable and have the risk of a single point of failure. Third, people have very weak or no control of their own data. This thesis explores the possibility of adopting the new distributed ledger technology, represented by blockchain, to address the problems. The relevant literature of blockchain and psychology tests was reviewed. A complete academic solution was proposed. It includes a permissioned blockchain, a no-SQL database, a web service, and a front-end. The blockchain stores user profile, metadata of psychology tests, final test scores, and access control data of the tests and test scores. The no-SQL database stores test materials and raw test results. The web service interacts with the blockchain and the no-SQL database. The front-end interacts with the web service. The solution was implemented, and the performance was evaluated. The evaluation results showed that the Post request is slower than the Get request and as the number of clients grows linearly, the latency of the requests grows linearly. The slower latency for the Post request compared to the Get request reflects the time it takes for the blockchain system to write information and change the common ledger status. The solution proposed here provides a new way to manage psychology test data with satisfactory performance. Future research can focus on extending the current solution to other questionnaire data management and to non-questionnaire-based psychology assessment data management.
Long-term prediction of multivariate time series is still an important but challenging problem. The key to solve this problem is to capture the spatial correlations at the same time, the spatio-temporal relationships at different times and the long-term dependence of the temporal relationships between different series. Attention-based recurrent neural networks (RNN) can effectively represent the dynamic spatio-temporal relationships between exogenous series and target series, but it only performs well in one-step time prediction and short-term time prediction. In this paper, inspired by human attention mechanism including the dual-stage two-phase (DSTP) model and the influence mechanism of target information and non-target information, we propose DSTP-based RNN (DSTP-RNN) and DSTP-RNN-2 respectively for long-term time series prediction. Specifically, we first propose the DSTP-based structure to enhance the spatial correlations between exogenous series. The first phase produces violent but decentralized response weight, while the second phase leads to stationary and concentrated response weight. Secondly, we employ multiple attentions on target series to boost the long-term dependence. Finally, we study the performance of deep spatial attention mechanism and provide experiment and interpretation. Our methods outperform nine baseline methods on four datasets in the fields of energy, finance, environment and medicine, respectively.
Cryptocurrencies are known as unpredictable due to their highly volatility. In time series, the forecasting accuracy is strongly affected by the methodologies that are used in identifying the pattern of a nonstationary stochastic realization. The purpose of the present study is to develop an algorithm that is capable of efficiently identifying the pattern of cryptocurrencies. A brief summary of the algorithm is given. To illustrate the quality of our proposed algorithm, we study the pattern of ten different reputable cryptocurrencies and use their daily closing prices to constitute a time series. The comparison between our proposed forecasting algorithm versus the autoregressive integrated moving average (ARIMA) process will be demonstrated.
Sep 8, 2017·Proceedings of the 2017 ACM International Joint Conference on Pervasive and Ubiquitous Computing and Proceedings of the 2017 ACM International Symposium on Wearable Computers
Modern IoT solutions are often an intricate system of interdependent components. Traditional monitoring techniques may not be sufficient to ensure the correct operation of those systems. We present an on-line machine learning approach for anomaly detection that is optimized for interpretability. The aim is to make it as intuitive as possible for human operators to derive insights about the system. To this end we combine characteristics of the system into sets of features that can be rendered graphically. Our solution builds on open source components and applies to any time series of numerical data. This work originated within a larger project on connected mobility that uses Blockchain technology to guarantee data integrity. Hence we demonstrate some results at the example of the public Ethereum blockchain. Further work will extend the solution to more general sensor data from the IoT realm.
Bitcoin is a cryptocurrency created by an unknown person or persons using the name of Satoshi Nakamoto. It is a decentralized, peer-to-peer currency. Even though all transactions are public, the identity of any particular party of a transaction is difficult to trace. Therefore, it is commonly used by criminals alongside legitimate users interested in privacy or novelty.
Discipline: Statistics
Faculty Mentor: Dr. Cristina Anton
Marco Stocchi, Maria Ilaria Lunesu, Simona Ibba, Gavina Baralla · 5 authors
In recent years search engines have become the go-to methods for achieving many types of knowledge, spanning from detailed descriptions or general information interesting to the user. Likewise several reassignment techniques are capturing the attention of researchers in the field of signal analysis. Particularly, the Synchrosqueezing Wavelet Transform - SST allows signal decomposition and instantaneous frequency extrusion, at the same time promising consistent reconstruction capabilities, hence the possibility to contrive an SST assisted inference engine. We are going to test it using datasets extracted from search engine trends, using a cloud of keywords related to the Bitcoin topic. This could be useful to study the evolution of the cryptocurrency both in time and geographical terms, and to estimate the future number of queries. The importance of Bitcoin queries prediction goes beyond the academic and research environments and, as such, it could lead to valuable commercial applications, such as financial recommender systems or blockchain-based transaction managers development.
Man-made self-organizing systems date back to antiquity; for example, elaborate water clocks found from Alexandria (Ctesibius's clepsydra) to Seoul (King Sejong's Chagyongnu) were designed to keep constant rates or strike at regular time intervals without human adjustment. More modern examples include Watt's centrifugal governor, Black's negative feedback amplifier, and Nyquist's stability test, which enabled engineered systems to stabilize themselves. These and numerous innovations in control theory and engineering optimization have contributed much to modern communications technology. But there is a need for advances in both methods and applications beyond what has been achieved in the precise settings of mechanical, electronic, and optical switching and transmission towards the self-management and control of large-scale systems with many interacting and semi-autonomous components. Natural phenomena may be a guide for us, a model for self-organized decentralized systems. Spontaneous magnetization, crystallization, lasers, and superconductivity are examples of structural self-organization in physics where cohesive behavior emerges from initial disorder. In self-assembly and auto-catalytic networks in chemistry, molecules organize themselves in well-ordered arrangements without external action, and in biology, we observe highly complex coordinated action as in the folding of proteins, homeostasis, and flocking.