Deep Transfer Learning for Time Series Data Based on Sensor Modality Classification

Frédéric Li; Kimiaki Shirahama; Muhammad Adeel Nisar; Xinyu Huang; Marcin Grzegorzek

doi:10.3390/s20154271

Deep Transfer Learning for Time Series Data Based on Sensor Modality Classification

Sensors ◽

10.3390/s20154271 ◽

2020 ◽

Vol 20 (15) ◽

pp. 4271 ◽

Cited By ~ 3

Author(s):

Frédéric Li ◽

Kimiaki Shirahama ◽

Muhammad Adeel Nisar ◽

Xinyu Huang ◽

Marcin Grzegorzek

Keyword(s):

Time Series ◽

Ubiquitous Computing ◽

Transfer Learning ◽

Time Series Data ◽

Wearable Devices ◽

Series Data ◽

Target Field ◽

Proper Training ◽

Application Fields ◽

General Architecture

The scarcity of labelled time-series data can hinder a proper training of deep learning models. This is especially relevant for the growing field of ubiquitous computing, where data coming from wearable devices have to be analysed using pattern recognition techniques to provide meaningful applications. To address this problem, we propose a transfer learning method based on attributing sensor modality labels to a large amount of time-series data collected from various application fields. Using these data, our method firstly trains a Deep Neural Network (DNN) that can learn general characteristics of time-series data, then transfers it to another DNN designed to solve a specific target problem. In addition, we propose a general architecture that can adapt the transferred DNN regardless of the sensors used in the target field making our approach in particular suitable for multichannel data. We test our method for two ubiquitous computing problems—Human Activity Recognition (HAR) and Emotion Recognition (ER)—and compare it a baseline training the DNN without using transfer learning. For HAR, we also introduce a new dataset, Cognitive Village-MSBand (CogAge), which contains data for 61 atomic activities acquired from three wearable devices (smartphone, smartwatch, and smartglasses). Our results show that our transfer learning approach outperforms the baseline for both HAR and ER.

Download Full-text

Intra-domain and cross-domain transfer learning for time series

10.5194/egusphere-egu21-12142 ◽

2021 ◽

Author(s):

Erik Otović ◽

Marko Njirjak ◽

Dario Jozinović ◽

Goran Mauša ◽

Alberto Michelini ◽

...

Keyword(s):

Machine Learning ◽

Time Series ◽

Transfer Learning ◽

Time Series Data ◽

The Other ◽

Series Data ◽

Sound Recognition ◽

Transfer Of Knowledge ◽

Learning Models ◽

Machine Learning Models

In this study, we compared the performance of machine learning models trained using transfer learning and those that were trained from scratch - on time series data. Four machine learning models were used for the experiment. Two models were taken from the field of seismology, and the other two are general-purpose models for working with time series data. The accuracy of selected models was systematically observed and analyzed when switching within the same domain of application (seismology), as well as between mutually different domains of application (seismology, speech, medicine, finance). In seismology, we used two databases of local earthquakes (one in counts, and the other with the instrument response removed) and a database of global earthquakes for predicting earthquake magnitude; other datasets targeted classifying spoken words (speech), predicting stock prices (finance) and classifying muscle movement from EMG signals (medicine). In practice, it is very demanding and sometimes impossible to collect datasets of tagged data large enough to successfully train a machine learning model. Therefore, in our experiment, we use reduced data sets of 1,500 and 9,000 data instances to mimic such conditions. Using the same scaled-down datasets, we trained two sets of machine learning models: those that used transfer learning for training and those that were trained from scratch. We compared the performances between pairs of models in order to draw conclusions about the utility of transfer learning. In order to confirm the validity of the obtained results, we repeated the experiments several times and applied statistical tests to confirm the significance of the results. The study shows when, within the set experimental framework, the transfer of knowledge brought improvements in terms of model accuracy and in terms of model convergence rate. Our results show that it is possible to achieve better performance and faster convergence by transferring knowledge from the domain of global earthquakes to the domain of local earthquakes; sometimes also vice versa. However, improvements in seismology can sometimes also be achieved by transferring knowledge from medical and audio domains. The results show that the transfer of knowledge between other domains brought even more significant improvements, compared to those within the field of seismology. For example, it has been shown that models in the field of sound recognition have achieved much better performance compared to classical models and that the domain of sound recognition is very compatible with knowledge from other domains. We came to similar conclusions for the domains of medicine and finance. Ultimately, the paper offers suggestions when transfer learning is useful, and the explanations offered can provide a good starting point for knowledge transfer using time series data.

Download Full-text

Multi-Channel Fusion Classification Method Based on Time-Series Data

Sensors ◽

10.3390/s21134391 ◽

2021 ◽

Vol 21 (13) ◽

pp. 4391

Author(s):

Xue-Bo Jin ◽

Aiqiang Yang ◽

Tingli Su ◽

Jian-Lei Kong ◽

Yuting Bai

Keyword(s):

Time Series ◽

Time Series Data ◽

Short Term Memory ◽

Evidence Theory ◽

Recurrence Plot ◽

Series Data ◽

Public Datasets ◽

Application Fields ◽

Original Time

Time-series data generally exists in many application fields, and the classification of time-series data is one of the important research directions in time-series data mining. In this paper, univariate time-series data are taken as the research object, deep learning and broad learning systems (BLSs) are the basic methods used to explore the classification of multi-modal time-series data features. Long short-term memory (LSTM), gated recurrent unit, and bidirectional LSTM networks are used to learn and test the original time-series data, and a Gramian angular field and recurrence plot are used to encode time-series data to images, and a BLS is employed for image learning and testing. Finally, to obtain the final classification results, Dempster–Shafer evidence theory (D–S evidence theory) is considered to fuse the probability outputs of the two categories. Through the testing of public datasets, the method proposed in this paper obtains competitive results, compensating for the deficiencies of using only time-series data or images for different types of datasets.

Download Full-text

Fuzzy Transfer Learning in Time Series Forecasting for Stock Market Prices

10.21203/rs.3.rs-1015226/v1 ◽

2021 ◽

Author(s):

Shanoli Samui Pal ◽

Samarjit Kar

Keyword(s):

Time Series ◽

Stock Market ◽

Prior Knowledge ◽

Transfer Learning ◽

Prediction Accuracy ◽

Time Series Data ◽

Time Series Prediction ◽

Series Data ◽

Data Sets ◽

Market Data

Abstract Transfer learning involves transferring prior knowledge of solving similar problems in order to achieve quick and efficient solution. The aim of fuzzy transfer learning is to transfer prior knowledge in an imprecise environment. Time series like stock market data are non-linear in nature and movement of stock is uncertain, so it is quite difficult following the stock market and in decision making. In this study, we propose a method to forecast stock market time series in the situation when we can use prior experience to make decisions. Fuzzy transfer learning (FuzzyTL) is based on knowledge transfer in that and adapting rules obtained domain. Three different stock market time series data sets are used for comparative study. It is observed that the effect of knowledge transferring works well together with smoothing of dependent attributes as the stock market data fluctuate with time. Finally, we give an empirical application in Shenzhen stock market with larger data sets to demonstrate the performance of the model. We have explored FuzzyTL in time series prediction to unerstand the essence of FuzzyTL. We were working on the question of the capability of FuzzyTL in improving prediction accuracy. From the comparisons, it can be said fuzzy transfer learning with smoothing improves prediction accuracy efficiently.

Download Full-text

A lightweight double-channel depthwise separable convolutional neural network for multimodal fusion gait recognition

Mathematical Biosciences and Engineering ◽

10.3934/mbe.2022055 ◽

2021 ◽

Vol 19 (2) ◽

pp. 1195-1212

Author(s):

Xiaoguang Liu ◽

◽

Meng Chen ◽

Tie Liang ◽

Cunguang Lou ◽

...

Keyword(s):

Neural Network ◽

Time Series ◽

Convolutional Neural Network ◽

Time Series Data ◽

Gait Recognition ◽

Wearable Devices ◽

Multimodal Fusion ◽

Model Complexity ◽

Series Data ◽

Double Channel

<abstract> Gait recognition is an emerging biometric technology that can be used to protect the privacy of wearable device owners. To improve the performance of the existing gait recognition method based on wearable devices and to reduce the memory size of the model and increase its robustness, a new identification method based on multimodal fusion of gait cycle data is proposed. In addition, to preserve the time-dependence and correlation of the data, we convert the time-series data into two-dimensional images using the Gramian angular field (GAF) algorithm. To address the problem of high model complexity in existing methods, we propose a lightweight double-channel depthwise separable convolutional neural network (DC-DSCNN) model for gait recognition for wearable devices. Specifically, the time series data of gait cycles and GAF images are first transferred to the upper and lower layers of the DC-DSCNN model. The gait features are then extracted with a three-layer depthwise separable convolutional neural network (DSCNN) module. Next, the extracted features are transferred to a softmax classifier to implement gait recognition. To evaluate the performance of the proposed method, the gait dataset of 24 subjects were collected. Experimental results show that the recognition accuracy of the DC-DSCNN algorithm is 99.58%, and the memory usage of the model is only 972 KB, which verifies that the proposed method can enable gait recognition for wearable devices with lower power consumption and higher real-time performance. </abstract>

Download Full-text

Towards Deep Industrial Transfer Learning for Anomaly Detection on Time Series Data

10.1109/etfa45728.2021.9613542 ◽

2021 ◽

Author(s):

Benjamin Maschler ◽

Tim Knodel ◽

Michael Weyrich

Keyword(s):

Time Series ◽

Anomaly Detection ◽

Transfer Learning ◽

Time Series Data ◽

Series Data ◽

Industrial Transfer

Download Full-text

Transfer Learning with Time Series Data: A Systematic Mapping Study

IEEE Access ◽

10.1109/access.2021.3134628 ◽

2021 ◽

pp. 1-1

Author(s):

Manuel Weber ◽

Maximilian Auch ◽

Christoph Doblander ◽

Peter Mandl ◽

Hans-Arno Jacobsen

Keyword(s):

Time Series ◽

Transfer Learning ◽

Time Series Data ◽

Series Data ◽

Systematic Mapping Study ◽

Mapping Study ◽

Systematic Mapping

Download Full-text

Intra-domain and cross-domain transfer learning for time series data – How transferable are the features?

Knowledge-Based Systems ◽

10.1016/j.knosys.2021.107976 ◽

2021 ◽

pp. 107976

Author(s):

Erik Otović ◽

Marko Njirjak ◽

Dario Jozinović ◽

Goran Mauša ◽

Alberto Michelini ◽

...

Keyword(s):

Time Series ◽

Transfer Learning ◽

Time Series Data ◽

Series Data ◽

Cross Domain ◽

Domain Transfer ◽

Domain Transfer Learning

Download Full-text

Graphical Exploratory Data Analysis for Categorical Longitudinal and Time Series Data

PsycEXTRA Dataset ◽

10.1037/e634372013-001 ◽

2013 ◽

Author(s):

Stephen J. Tueller ◽

Richard A. Van Dorn ◽

Georgiy Bobashev ◽

Barry Eggleston

Keyword(s):

Time Series ◽

Data Analysis ◽

Exploratory Data Analysis ◽

Time Series Data ◽

Series Data ◽

Exploratory Data

Download Full-text

Faktor-Faktor Yang Mempengaruhi Nilai Tukar Dollar Amerika Serikat Terhadap Rupiah Tahun 2000–2013

Jurnal Riset Manajemen Sekolah Tinggi Ilmu Ekonomi Widya Wiwaha Program Magister Manajemen ◽

10.32477/jrm.v1i2.72 ◽

2017 ◽

Vol 1 (2) ◽

pp. 177-191

Author(s):

Rizki Rahma Kusumadewi ◽

Wahyu Widayat

Keyword(s):

Time Series ◽

Exchange Rate ◽

Money Supply ◽

Time Series Data ◽

The United States ◽

Economic Conditions ◽

Series Data ◽

Arch Model ◽

United States Dollar ◽

The Exchange Rate

Exchange rate is one tool to measure a country’s economic conditions. The growth of a stable currency value indicates that the country has a relatively good economic conditions or stable. This study has the purpose to analyze the factors that affect the exchange rate of the Indonesian Rupiah against the United States Dollar in the period of 2000-2013. The data used in this study is a secondary data which are time series data, made up of exports, imports, inflation, the BI rate, Gross Domestic Product (GDP), and the money supply (M1) in the quarter base, from first quarter on 2000 to fourth quarter on 2013. Regression model time series data used the ARCH-GARCH with ARCH model selection indicates that the variables that significantly influence the exchange rate are exports, inflation, the central bank rate and the money supply (M1). Whereas import and GDP did not give any influence.

Download Full-text