Cryptocurrency forecasting with deep learning chaotic neural networks
پیش بینی cryptocurrency با یادگیری عمیق شبکه های عصبی پر هرج و مرج-2019
We implement deep learning techniques to forecast the price of the three most widely traded digital currencies i.e., Bitcoin, Digital Cash and Ripple. To the best of our knowledge, this is the first work to make use of deep learning in cryptocurrency prediction. The results from testing the existence of non- linearity revealed that the time series of all digital currencies exhibit fractal dynamics, long memory and self-similarity. The predictability of long-short term memory neural network topologies (LSTM) is signif- icantly higher when compared to the generalized regression neural architecture, set forth as our bench- mark system. The latter failed to approximate global nonlinear hidden patterns regardless of the degree of contamination with noise, as they are based on Gaussian kernels suitable only for local approximation of non-stationary signals. Although the computational burden of the LSTM model is higher as opposed to brute force in nonlinear pattern recognition, eventually deep learning was found to be highly efficient in forecasting the inherent chaotic dynamics of cryptocurrency markets.
Keywords: Digital currencies | Deep learning | Fractality | Neural networks | Chaos | Forecasting
Data-based structure selection for unified discrete grey prediction model
Data-based structure selection for unified discrete grey prediction model-2019
Grey models have been reported to be promising for time series prediction with small samples, but the diversity kinds of model structures and modelling assumptions restrains their further applications and developments. In this paper, a novel grey prediction model, named discrete grey polynomial model, is proposed to unify a family of univariate discrete grey models. The proposed model has the capacity to represent most popular homogeneous and non-homogeneous discrete grey models and furthermore, it can induce some other novel models, thereby highlighting the relationship between the models and their structures and assumptions. Based on the proposed model, a data-based algorithm is put forward to se- lect the model structure adaptively. It reduces the requirement for modeler’s knowledge from an expert system perspective. Two numerical experiments with large-scale simulations are conducted and the re- sults show its effectiveness. In the end, two real case tests show that the proposed model benefits from its adaptive structure and produces reliable multi-step ahead predictions.
Keywords: Grey system theory | Discrete grey model | Structure selection | Matrix decomposition
Improving air quality prediction accuracy at larger temporal resolutions using deep learning and transfer learning techniques
بهبود دقت پیش بینی کیفیت هوا در وضوح زمانی بزرگتر با استفاده از تکنیک های یادگیری عمیق و انتقال یادگیری-2019
As air pollution becomes more and more severe, air quality prediction has become an important approach for air pollution management and prevention. In recent years, a number of methods have been proposed to predict air quality, such as deterministic methods, statistical methods as well as machine learning methods. However, these methods have some limitations. Deterministic methods require expensive computations and specific knowledge for parameter identification, while the forecasting performance of statistical methods is limited due to the linear assumption and the multicollinearity problem. Most of the machine learning methods, on the other hand, cannot capture the time series patterns or learn from the long-term dependencies of air pollutant concentrations. Furthermore, there is a lack of methods that could generate high prediction accuracy for air quality forecasting at larger temporal resolutions, such as daily and weekly or even monthly. This paper, therefore, proposes a deep learning-based method namely transferred bi-directional long short-term memory (TL-BLSTM) model for air quality prediction. The methodology framework utilizes the bi-directional LSTM model to learn from the longterm dependencies of PM2.5, and applies transfer learning to transfer the knowledge learned from smaller temporal resolutions to larger temporal resolutions. A case study is conducted in Guangdong, China to test the proposed methodology framework. The performance of the framework is compared with other commonly seen machine learning algorithms, and the results show that the proposed TL-BLSTM model has smaller errors, especially for larger temporal resolutions
Keywords: Air quality prediction | Large temporal resolution | Deep learning | Long short-term memory | Transfer learning
HMM-based Supervised Machine Learning Framework for the Detection of ECG R Peak Locations
چارچوب یادگیری ماشین نظارت شده مبتنی بر HMM مبتنی برای تشخیص مکان های اوج ECG-2019
Objective: Fetal Electro Cardiogram (fECG) provides critical information on the wellbeing of a foetus heart in its developing stages in the mother’s womb. The objective of this work is to extract fECG which is buried in a composite signal consisting of itself, maternal ECG (mECG) and noises contributed from various unavoidable sources. In the past, the challenge of extracting fECG from the composite signal was dealt with by Stochastic Weiner filter, model-based Kalman filter and other adaptive filtering techniques. Blind Source Separation (BSS) based Independent Component Analysis (ICA) has shown an edge over the adaptive filtering techniques as the former does not require a reference signal. Recently, data-driven machine learning techniques e.g., adaptive neural networks, adaptive neuro-fuzzy inference system, support vector machine (SVM) are also applied. Method: This work pursues hidden Markov model (HMM)-based supervised machine learning frame-work for the determination of the location of fECG QRS complex from the composite abdominal signal. HMM is used to model the underlying hidden states of the observable time series of the extracted and separated fECG data with its QRS peak location as one of the hidden states. The state transition probabilities are estimated in the training phase using the annotated data sets. Afterwards, using the estimated HMM networks, fQRS locations are detected in the testing phase. To evaluate the proposed technique, the accuracy of the correct detection of QRS complex with respect to the correct annotation of QRS complex location is considered and quantified by the sensitivity, probability of false alarm, and accuracy. Results: The best results that have been achieved using the proposed method are: accuracy – 97.1%, correct detection rate (translated to sensitivity) – 100%, and false alarm rate – 2.89%.
Keywords: fECG | mECG | Machine learning | HMM | Accuracy | Sensitivity
Mining Twitter data for causal links between tweets and real-world outcomes
استخراج داده های توییتر برای پیوندهای علی بین توییتها و پیامدهای دنیای واقعی-2019
The authors present an expert and intelligent system that (1) identifies influential term groups having causal relationships with real-world enterprise outcomes from Twitter data and (2) quantifies the appro- priate time lags between identified influential term groups and enterprise outcomes. Existing expert and intelligent systems, which are defined as computer systems that imitate the ability of human decision making, could enable computers to identify the spread of Twitter users’ enterprise-related feedback au- tomatically. However, existing expert and intelligent systems have limitations on automatically identifying the causal effects on enterprise outcomes. Identifying the causal effects on enterprise outcomes is impor- tant, because Twitter users’ feedback toward enterprise decisions may have real-world implications. The proposed expert and intelligent system can support decision makers’ decisions considering the real-world effects of identified Twitter users’ feedback on enterprise outcomes. In particular, (1) a co-occurrence net- work analysis model is exploited to discover term candidates for generating influential term groups that are combinations of enterprise-related terms, which potentially influence enterprise outcomes. (2) Time series models and (3) a Granger causality analysis model are then employed to identify influential term groups having causal relationships with enterprise outcomes with the appropriate time lags. Case studies involving a real-world internet video streaming and disc rental provider as well as an airline company are used to test the validity of the proposed expert and intelligent system for both predicting enterprise outcomes in a long period and predicting the effects of specific events on enterprise outcomes in a short period.
Keywords: Expert and intelligent system | Social media | Enterprise outcome | Co-occurrence network | Time series analysis | Granger causality analysis
Deep learning approach for sustainable WWTP operation: A case study on data-driven influent conditions monitoring
رویکرد یادگیری عمیق برای عملکرد پایدار WWTP: مطالعه موردی در زمینه نظارت بر شرایط تأثیرگذار محور داده-2019
Wastewater treatment plants (WWTPs) are sustainable solutions to water scarcity. As initial conditions offered to WWTPs, influent conditions (ICs) affect treatment units states, ongoing processes mechanisms, and product qualities. Anomalies in ICs, often raised by abnormal events, need to be monitored and detected promptly to improve system resilience and provide smart environments. This paper proposed and verified data-driven anomaly detection approaches based on deep learning methods and clustering algorithms. Combining both the ability to capture temporal auto-correlation features among multivariate time series from recurrent neural networks (RNNs), and the function to delineate complex distributions from restricted Boltzmann machines (RBM), RNN-RBM models were employed and connected with various classifiers for anomaly detection. The effectiveness of RNN based, RBM based, RNN-RBM based, or standalone individual detectors, including expectation maximization clustering, K-means clustering, mean-shift clustering, one-class support vector machine (OCSVM), spectral clustering, and agglomerative clustering algorithms were evaluated by importing seven years ICs data from a coastal municipal WWTP where more than 150 abnormal events occurred. Results demonstrated that RNN-RBM-based OCSVM approach outperformed all other scenarios with an area under the curve value up to 0.98, which validated the superiority in feature extraction by RNN-RBM, and the robustness in multivariate nonlinear kernels by OCSVM. The model was flexible for not requiring assumptions on data distribution, and could be shared and transferred among environmental data scientists.
Keywords: Wastewater treatment plant | Influent conditions monitoring | Machine learning | Unsupervised deep learning
Using machine learning to quantify the impacts of genetically modified crops on US midwest corn yields
استفاده از یادگیری ماشینی برای تعیین کمیت تأثیر محصولات اصلاح شده ژنتیکی بر عملکرد ذرت میان غربی ایالات متحده-2019
Global food security is becoming increasingly stressed by growing populations and climate change. To compensate for these stresses, crop yields must increase throughout the upcoming century. One of the more prominently featured solutions entails genetically modified crops, but their impacts on yields are contested. Here, we leverage machine learning techniques to examine the effects genetically modified crops have had on US corn yields. In particular, a principal components analysis conducted on US Midwest county yields reveals that the commercialization of genetically modified corn accentuated preexisting spatial disparities in production and explains approximately 6–12% of the regions inter-county variation in yields from 1980 to 2015. Additionally, counterfactual yield trajectories predicted by Bayesian structural time series models using non-genetically modified crops as synthetic controls suggest that the adoption of this biotechnology amounted to an approximate 13% increase in overall US corn yields from 1996 to 2015.
Keywords: GM crops | Principal components analysis | Corn yields | Machine learning | Bayesian structural time series
DOSED: A deep learning approach to detect multiple sleep micro-events in EEG signal
DOSED: یک رویکرد یادگیری عمیق برای تشخیص ریز وقایع چند خوابی در سیگنال EEG-2019
Background: Electroencephalography (EEG) monitors brain activity during sleep and is used to identify sleep disorders. In sleep medicine, clinicians interpret raw EEG signals in so-called sleep stages, which are assigned by experts to every 30s window of signal. For diagnosis, they also rely on shorter prototypical micro-architecture events which exhibit variable durations and shapes, such as spindles, K-complexes or arousals. Annotating such events is traditionally performed by a trained sleep expert, making the process time consuming, tedious and subject to inter-scorer variability. To automate this procedure, various methods have been developed, yet these are event-specific and rely on the extraction of hand-crafted features. New method: We propose a novel deep learning architecture called Dreem One Shot Event Detector (DOSED). DOSED jointly predicts locations, durations and types of events in EEG time series. The proposed approach, applied here on sleep related micro-architecture events, is inspired by object detectors developed for computer vision such as YOLO and SSD. It relies on a convolutional neural network that builds a feature representation from raw EEG signals, as well as two modules performing localization and classification respectively. Results and comparison with other methods: The proposed approach is tested on 4 datasets and 3 types of events (spindles, K-complexes, arousals) and compared to the current state-of-the-art detection algorithms. Conclusions: Results demonstrate the versatility of this new approach and improved performance compared to the current state-of-the-art detection methods.
Keywords: Deep learning | Machine learning | EEG | Event detection | Sleep
Deep learning in exchange markets
یادگیری عمیق در بازارهای ارز-2019
We present the implementation of a short-term forecasting system of price movements in exchange mar- kets using market depth data and a systematic procedure to enable a fully automated trading system. Three types of Deep Learning (DL) Neural Network (NN) methodologies are trained and tested: Deep NN Classifier (DNNC), Long Short-Term Memory (LSTM) and Convolutional NN (CNN). Although the LSTM is more suitable for multivariate time series analysis from a theoretical point of view, test results indicate that the CNN has on average the best predictive power in the case study under analysis, which is the UK to Win Horse Racing market during pre-live stage in the world’s most relevant betting exchange. Implica- tions from the generalized use of automated trading systems in betting exchange markets are discussed.
Keywords: Deep learning | Betting exchange | Market depth | Classification
Machine Learning based Digital Twin Framework for Production Optimization in Petrochemical Industry
چارچوب دوقلوی دیجیتال مبتنی بر یادگیری ماشین برای بهینه سازی تولید در صنعت پتروشیمی-2019
Digital twins, along with the internet of things (IoT), data mining, and machine learning technologies, offer great potential in the transformation of today’s manufacturing paradigm toward intelligent manufacturing. Production control in petrochemical industry involves complex circumstances and a high demand for timeliness; therefore, agile and smart controls are important components of intelligent manufacturing in the petrochemical industry. This paper proposes a framework and approaches for constructing a digital twin based on the petrochemical industrial IoT, machine learning and a practice loop for information exchange between the physical factory and a virtual digital twin model to realize production control optimization. Unlike traditional production control approaches, this novel approach integrates machine learning and real-time industrial big data to train and optimize digital twin models. It can support petrochemical and other process manufacturing industries to dynamically adapt to the changing environment, respond in a timely manner to changes in the market due to production optimization, and improve economic benefits. Accounting for environmental characteristics, this paper provides concrete solutions for machine learning difficulties in the petrochemical industry, e.g., high data dimensions, time lags and alignment between time series data, and high demand for immediacy. The approaches were evaluated by applying them in the production unit of a petrochemical factory, and a model was trained via industrial IoT data and used to realize intelligent production control based on real-time data. A case study shows the effectiveness of this approach in the petrochemical industry.
Keywords: digital twin | machine learning | internet of things | petrochemical industry | production control optimization