Data-driven Machine Learning Algorithms Research Articles

A data-driven machine learning algorithm for financial market prediction

A data-driven machine learning algorithm for financial market prediction

Machine learning approaches can reduce environmental data requirements for regional yield potential simulation

The crop growth model (CGM) widely contributes to studying the impacts of regional climate change on crop growth status and yield. However, it requires high-quality daily weather data for the model establishment and verification, and the low data availability limits the application of CGM at the regional scale. With the rapid development of machine learning techniques, adapting data-driven machine learning algorithms to build a meta-model with environmental feature variables at the desired spatio-temporal scale provides a new method for regional yield simulation. In this study, we developed four machine-learning-based meta-models to simulate regional yield potential (RYP) of wheat in China with the selected environmental feature variables and assessed four different machine learning algorithms, including multiple linear regression (MLR), artificial neural networks (ANN), random forest (RF) and support vector regression (SVR). The research aimed to verify whether these meta-models can outperform CGM in simulating RYP. The results showed that the meta-models could reduce the requirements of the number of input variables and the amount of data for RYP simulation and maintain the simulation accuracy because monthly weather variables could replace daily weather variables in the meta-models. Although all four meta-models can well reveal the mean RYP, the meta-model based on RF has the best performance. In the RF-based meta-model, longitude, latitude, altitude, and the averaged maximum temperature in March are the top four ranked essential variables. However, the generalizability of meta-models is affected by the training dataset, and the meta-models cannot adapt appropriately to new unseen data. Moreover, there is no unique optimal machine learning algorithm used for building meta-models, and this will increase the workload of similar research in the future.

Solid-State Lithium Battery Cycle Life Prediction Using Machine Learning

Battery lifetime prediction is a promising direction for the development of next-generation smart energy storage systems. However, complicated degradation mechanisms, different assembly processes, and various operation conditions of the batteries bring tremendous challenges to battery life prediction. In this work, charge/discharge data of 12 solid-state lithium polymer batteries were collected with cycle lives ranging from 71 to 213 cycles. The remaining useful life of these batteries was predicted by using a machine learning algorithm, called symbolic regression. After populations of breed, mutation, and evolution training, the test accuracy of the quantitative prediction of cycle life reached 87.9%. This study shows the great prospect of a data-driven machine learning algorithm in the prediction of solid-state battery lifetimes, and it provides a new approach for the batch classification, echelon utilization, and recycling of batteries.

BOOSTING SEGMENTATION ACCURACY OF THE DEEP LEARNING MODELS BASED ON THE SYNTHETIC DATA GENERATION

Abstract. In the era of data-driven machine learning algorithms, data represents a new oil. The application of machine learning algorithms shows they need large heterogeneous datasets that crucially are correctly labeled. However, data collection and its labeling are time-consuming and labor-intensive processes. A particular task we solve using machine learning is related to the segmentation of medical devices in echocardiographic images during minimally invasive surgery. However, the lack of data motivated us to develop an algorithm generating synthetic samples based on real datasets. The concept of this algorithm is to place a medical device (catheter) in an empty cavity of an anatomical structure, for example, in a heart chamber, and then transform it. To create random transformations of the catheter, the algorithm uses a coordinate system that uniquely identifies each point regardless of the bend and the shape of the object. It is proposed to take a cylindrical coordinate system as a basis, modifying it by replacing the Z-axis with a spline along which the h-coordinate is measured. Having used the proposed algorithm, we generated new images with the catheter inserted into different heart cavities while varying its location and shape. Afterward, we compared the results of deep neural networks trained on the datasets comprised of real and synthetic data. The network trained on both real and synthetic datasets performed more accurate segmentation than the model trained only on real data. For instance, modified U-net trained on combined datasets performed segmentation with the Dice similarity coefficient of 92.6±2.2%, while the same model trained only on real samples achieved the level of 86.5±3.6%. Using a synthetic dataset allowed decreasing the accuracy spread and improving the generalization of the model. It is worth noting that the proposed algorithm allows reducing subjectivity, minimizing the labeling routine, increasing the number of samples, and improving the heterogeneity.

Comparative performance analysis of support vector regression and artificial neural network for prediction of municipal solid waste generation.

The evolution of machine learning (ML) algorithms provides researchers and engineers with state-of-the-art tools to dynamically model complex relationships. The design and operation of municipal solid waste (MSW) management systems require accurate estimation of generation rates. In this study, we applied rapid, non-linear and non-parametric data driven ML algorithms independently, multi-layer perceptron artificial neural network (MLP-ANN) and support vector regression (SVR) models to predict annual MSW generation rates in Bahrain. Models were trained and tested with MSW generation data for period of 1997-2019. The population, gross domestic product, annual tourist numbers, annual electricity consumption and total annual CO2 emissions were selected as explanatory variables and incorporated into developed models. The zero score normalization (ZSN) and minimum maximum normalization (MMN) methods were utilized to improve the quality of data and subsequently enhances the performance of ML algorithms. Statistical metrics were employed to discriminate performance of MLP-ANN and SVR models. The linear, polynomial, radial basis function (RBF) and sigmoid kernel functions were investigated to find the optimal SVR model. Results showed that RBF-SVR model with R2 value of 0.97% and 4.82% and absolute forecasting error (AFE) for the period of 2008 and 2019 exhibits superior prediction and robustness in comparison to MLP-ANN. The efficacy of MLP-ANN model was also reasonably successful with R2 value of 0.94. It was shown that MMN pre-processing generated optimal MLP-ANN model while ZSN pre-processing produced optimal RBF-SVR model. This work also highlights the importance of application of ML modelling approaches to plan and implement their roadmap for waste management systems by policymakers.

Imputation of missing well log data by random forest and its uncertainty analysis

Well logs are commonly used by geoscientists to infer and extrapolate physical properties of subsurface rocks. However, at some depth intervals, well log values might be missing due to operational issues in the logging process. To overcome this problem, an innovative approach to reconstruct well logs is proposed using machine learning methods. Based on other complete logging features, the missing well log values are predicted by data-driven machine learning algorithms, namely random forest. A grid-searching scheme is applied to find a combination of hyper-parameters for the best cross-validation score. During the training process, the relative importance of different input features is analysed to remove weakly sensitive measurements and prioritize data with strong correlation with the target variables. Principal component analysis is applied to explore the multicollinearity in the input features, such that only few principal components in the new data vector are used to represent a large fraction of the variance in the original data. To quantify the uncertainty in the predictions, a quantile regression tree is used for determining prediction intervals. Well log data from the Volve Field are used for validation of the prediction obtained by random forest, in which a high correlation coefficient between prediction and reference is achieved. The prediction intervals of different percentiles are estimated, and show more accurate results at depth points where a small range of the prediction intervals exists.

Transient Prediction of Nanoparticle-Laden Droplet Drying Patterns through Dynamic Mode Decomposition

Nanoparticle-laden sessile droplet drying has a wide impact on applications. However, the complexity affected by the droplet evaporation dynamics and particle self-assembly behavior leads to challenges in the accurate prediction of the drying patterns. We initiate a data-driven machine learning algorithm by using a single data collection point via a top-view camera to predict the transient drying patterns of aluminum oxide (Al2O3) nanoparticle-laden sessile droplets with three cases according to particle sizes of 5 and 40 nm and Al2O3 concentrations of 0.1 and 0.2 wt %. Dynamic mode decomposition is used as the data-driven learning model to recognize each nanoparticle-laden droplet as an individual system and then apply the transfer learning procedure. Along 270 s of droplet drying experiments, the training period of the first 100 s is selected, and then the rest of the 170 s is predicted with less than a 10% error between the predicted and the actual droplet images. The developed data-driven approach has also achieved the acceptable prediction for the droplet diameter with less than 0.13% error and a coffee-ring thickness over a range of 2.0 to 6.7 μm. Moreover, the proposed machine learning algorithm can recognize the volume of the droplet liquid and the transition of the drying regime from one to another according to the predicted contact line and the droplet height.

Lamb wave based damage detection in metallic plates using multi-headed 1-dimensional convolutional neural network

Lamb wave based damage diagnosis holds potential for real-time structural health monitoring; however, analysing the Lamb wave response possess challenge due to its complex physics. Data-driven machine learning (ML) algorithms are often more effective in identifying the damage-related features from these complex responses. However, in analysing such complex responses the ML algorithms requires extensive data pre-processing and are often not suitable for real-time damage detection. This paper presents a deep learning multi-headed 1-dimensional convolutional neural network (1D-CNN) architecture capable to operate directly on raw discrete time-domain Lamb wave signals recorded from a thin metallic plate. The multi-headed configuration consisting of two parallel 1D-CNN layers is capable to learn higher order damage-related features and enhances robustness of overall classification performance. To train the adopted 1D-CNN algorithm a diverse database is also constructed consisting 216 numerically and 24 experimentally generated responses of a thin 1.6 mm Al-5052 plate structure. The diversification of training database is achieved by varying parameters like scanning length, scanning frequency and adding different levels of white noises to the captured responses. Later, the trained 1D-CNN architecture is tested against two separated unseen test-databases. The first test database consist of experimentally generated 12 samples with notch-like damage and 12 samples of pristine condition. The proposed 1D-CNN classifier generalizes well on the unseen samples and decisively predicts the outcome for 23 out of 24 samples of first test database. The second test database consists of 108 unseen FE simulated samples capturing additional damage scenarios. In the second test phase, the model has correctly predicted the condition of all the 108 samples.

The Research of SEIJR Model With Time-Delay Based on 2019-nCov

A global epidemic disease known as the novel coronavirus (2019-nCov) had seriously hit the most area around the whole world causing unpredictable loss of manpower and finance during the past year. Modeling the spread and development of infectious diseases represented by new Coronavirus has become an important part of public health work in the world. Estimation of possible infection population and prospective suggestion of handling spread based on existing data is crucial. In this article, we build a more applicable model called SEIJR with a log-normal distributed time delay to forecast the trend of spreading considering the biology parameters obtained based on Chinese clinical data in Wuhan and the real spread feature of 2019-nCov in Italy. Adopting Particle Swarm Optimization (PSO), we estimate the early period average spreading velocity ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\boldsymbol {\alpha _{0}}$ </tex-math></inline-formula> ) and implement inversion analysis of time point ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\boldsymbol {T_{0}}$ </tex-math></inline-formula> ) when the virus first hit Italy. Based on fixed <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\boldsymbol {\alpha _{0}}$ </tex-math></inline-formula> and <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\boldsymbol {T_{0}}$ </tex-math></inline-formula> , we then obtained the average spreading velocity <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\boldsymbol {\alpha _{1}}$ </tex-math></inline-formula> after the area lockdown using PSO. The result shows that it will help address the infection by generating the prediction trends of different <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\boldsymbol {\alpha }$ </tex-math></inline-formula> which we considered. Finally, our research applies Logistic regression, Neural Network embedding LSTM layer, which is two representative machine learning algorithms, to directly predict future infection trends and compare the forecast with results yielded by mathematical model adopting differential equations. Not only solved the complex, nondifferentiable equation of the epidemic model, this research also performs well in inversion analysis based on PSO which conveys informative outcomes for further discussion on precautious action. The comparison with the machine learning algorithms shows that the 2019-nCov based epidemic dynamics assumption is reasonable and helpful to the mathematical model, which is better than the data-driven machine learning algorithms. Code can be freely downloaded from <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/Summerwork/2019-nCov-Prediction</uri> .

Prediction of EV Charging Behavior Using Machine Learning

As a key pillar of smart transportation in smart city applications, electric vehicles (EVs) are becoming increasingly popular for their contribution in reducing greenhouse gas emissions. One of the key challenges, however, is the strain on power grid infrastructure that comes with large-scale EV deployment. The solution to this lies in utilization of smart scheduling algorithms to manage the growing public charging demand. Using data-driven tools and machine learning algorithms to learn the EV charging behavior can improve scheduling algorithms. Researchers have focused on using historical charging data for predictions of behavior such as departure time and energy needs. However, variables such as weather, traffic, and nearby events, which have been neglected to a large extent, can perhaps add meaningful representations, and provide better predictions. Therefore, in this paper we propose the usage of historical charging data in conjunction with weather, traffic, and events data to predict EV session duration and energy consumption using popular machine learning algorithms including random forest, SVM, XGBoost and deep neural networks. The best predictive performance is achieved by an ensemble learning model, with SMAPE scores of 9.9% and 11.6% for session duration and energy consumptions, respectively, which improves upon the existing works in the literature. In both predictions, we demonstrate a significant improvement compared to previous work on the same dataset and we highlight the importance of traffic and weather information for charging behavior predictions.

Failure risk analysis of pipelines using data-driven machine learning algorithms

Failure risk analysis of pipeline networks is essential for their effective management. Since pipeline networks are often large and complex, analyzing a large number of pipelines is often intensive and time-consuming. This study utilizes recent advances in machine learning to develop a viable alternative to computationally intensive analytical approaches to determine the failure risk of steel oil and gas pipelines. Burst failure risk of pipelines under active corrosion defects is evaluated considering the remaining strength of pipelines with corrosion pit. In the first part of the study, a comprehensive database of pipelines is generated based on information extracted from literature, and true failure pressure is predicted considering variability between predicted burst failure pressure and experimental burst test results. True burst failure pressure is estimated using various design codes in a probabilistic manner. Among the various burst failure prediction models, the DNV RP-F101 model provides the lowest variability in the failure pressure prediction. Therefore, the model is used to generate the probability of failure of pipelines at a burst limit state. Pipelines are classified based on their probability of failures, from low to severe failure risk groups. Then, eight machine learning algorithms are evaluated based on the generated dataset to identify the best failure prediction model. The performance of machine learning algorithms is evaluated on the basis of a confusion matrix and computational efficiency. It shows that XGBoost is the optimal algorithm in predicting the failure and is recommended for future analysis. The computational efficiency of the machine learning algorithm is also compared to the physics-based model. It is revealed that machine learning algorithms can perform a failure risk analysis of pipelines with greater computational efficiency than the physics-based approach. Even the slowest machine learning algorithm is about 12 times faster than the physics-based model.

Contour Stella Image and Deep Learning for Signal Recognition in the Physical Layer

The rapid development of communication systems poses unprecedented challenges, e.g., handling exploding wireless signals in a real-time and fine-grained manner. Recent advances in data-driven machine learning algorithms, especially deep learning (DL), show great potential to address the challenges. However, waveforms in the physical layer may not be suitable for the prevalent classical DL models, such as convolution neural network (CNN) and recurrent neural network (RNN), which mainly accept formats of images, time series, and text data in the application layer. Therefore, it is of considerable interest to bridge the gap between signal waveforms to DL amenable data formats. In this article, we develop a framework to transform complex-valued signal waveforms into images with statistical significance, termed contour stellar image (CSI), which can convey deep level statistical information from the raw wireless signal waveforms while being represented in an image data format. In this article, we explore several potential application scenarios and present effective CSI-based solutions to address the signal recognition challenges. Our investigation validates that CSI is a promising method to bridge the gap between signal recognition and DL.

Product Completion Time Prediction Using A Hybrid Approach Combining Deep Learning and System Model

The prediction of product completion time is critical in real time production scheduling and control to achieve customer demand satisfaction. However, it is a challenging task due to the increasing complexity of production systems and greater diversity of products. The recent advancement in data-driven approach and machine learning algorithms have provided unprecedent opportunities to tackle such problems that otherwise very difficult to solve using conventional methods in the manufacturing industry. However, most existing studies on product completion time prediction adopt a purely data-driven approach while ignoring the prospect of integrating domain knowledge in their machine learning models. In this paper, we propose a hybrid approach to predicting product completion time by combining the strengths of both machine learning techniques and analytical system model. A mathematical model for multi-product serial production line is proposed to describe the real-time dynamics of the system. With this model, the strict lower bound of product completion time can be efficiently computed for given system status, where the lower bound represents the least possible product completion time when assuming no random downtime in the system. Instead of directly predicting the product completion time, a deep learning model is developed to only predict the distance between the lower bound and actual product completion time. Guided by properties of production system, we discover a recurrent sequence in the prediction problem by modeling each machine and product as recurrent units. The Long Short-Term Memory (LSTM) method, a prominent variant of recurrent neural network (RNN), is used to combine with the system model to predict the product completion time in a real-time fashion.

Is Clustering Time-Series Water Depth Useful? An Exploratory Study for Flooding Detection in Urban Drainage Systems

As sensor measurements emerge in urban water systems, data-driven unsupervised machine learning algorithms have drawn tremendous interest in event detection and hydraulic water level and flow prediction recently. However, most of them are applied in water distribution systems and few studies consider using unsupervised cluster analysis to group the time-series hydraulic-hydrologic data in stormwater urban drainage systems. To improve the understanding of how cluster analysis contributes to flooding location detection, this study compared the performance of K-means clustering, agglomerative clustering, and spectral clustering in uncovering time-series water depth dissimilarity. In this work, the water depth datasets are simulated by an urban drainage model and then formatted for a clustering problem. Three standard performance evaluation metrics, namely the silhouette coefficient index, Calinski–Harabasz index, and Davies–Bouldin index are employed to assess the clustering performance in flooding detection under various storms. The results show that silhouette coefficient index and Davies–Bouldin index are more suitable for assessing the performance of K-means and agglomerative clustering, while the Calinski–Harabasz index only works for spectral clustering, indicating these clustering algorithms are metric-dependent flooding indicators. The results also reveal that the agglomerative clustering performs better in detecting short-duration events while K-means and spectral clustering behave better in detecting long-duration floods. The findings of these investigations can be employed in urban stormwater flood detection at the specific junction-level sites by using the occurrence of anomalous changes in water level of correlated clusters as flood early warning for the local neighborhoods.

ScCancer: a package for automated processing of single-cell RNA-seq data in cancer.

Molecular heterogeneities and complex microenvironments bring great challenges for cancer diagnosis and treatment. Recent advances in single-cell RNA-sequencing (scRNA-seq) technology make it possible to study cancer cell heterogeneities and microenvironments at single-cell transcriptomic level. Here, we develop an R package named scCancer, which focuses on processing and analyzing scRNA-seq data for cancer research. Except basic data processing steps, this package takes several special considerations for cancer-specific features. Firstly, the package introduced comprehensive quality control metrics. Secondly, it used a data-driven machine learning algorithm to accurately identify major cancer microenvironment cell populations. Thirdly, it estimated a malignancy score to classify malignant (cancerous) and non-malignant cells. Then, it analyzed intra-tumor heterogeneities by key cellular phenotypes (such as cell cycle and stemness), gene signatures and cell-cell interactions. Besides, it provided multi-sample data integration analysis with different batch-effect correction strategies. Finally, user-friendly graphic reports were generated for all the analyses. By testing on 56 samples with 433405 cells in total, we demonstrated its good performance. The package is available at: http://lifeome.net/software/sccancer/.

Data Augmentation and Pretraining for Template-Based Retrosynthetic Prediction in Computer-Aided Synthesis Planning.

This work presents efforts to augment the performance of data-driven machine learning algorithms for reaction template recommendation used in computer-aided synthesis planning software. Often, machine learning models designed to perform the task of prioritizing reaction templates or molecular transformations are focused on reporting high-accuracy metrics for the one-to-one mapping of product molecules in reaction databases to the template extracted from the recorded reaction. The available templates that get selected for inclusion in these machine learning models have been previously limited to those that appear frequently in the reaction databases and exclude potentially useful transformations. By augmenting open-access data sets of organic reactions with explicitly calculated template applicability and pretraining a template-relevance neural network on this augmented applicability data set, we report an increase in the template applicability recall and an increase in the diversity of predicted precursors. The augmentation and pretraining effectively teaches the neural network an increased set of templates that could theoretically lead to successful reactions for a given target. Even on a small data set of well-curated reactions, the data augmentation and pretraining methods resulted in an increase in top-1 accuracy, especially for rare templates, indicating that these strategies can be very useful for small data sets.

An intelligent approach for data pre-processing and analysis in predictive maintenance with an industrial case study

Recent development in the predictive maintenance field has focused on incorporating artificial intelligence techniques in the monitoring and prognostics of machine health. The current predictive maintenance applications in manufacturing are now more dependent on data-driven Machine Learning algorithms requiring an intelligent and effective analysis of a large amount of historical and real-time data coming from multiple streams (sensors and computer systems) across multiple machines. Therefore, this article addresses issues of data pre-processing that have a significant impact on generalization performance of a Machine Learning algorithm. We present an intelligent approach using unsupervised Machine Learning techniques for data pre-processing and analysis in predictive maintenance to achieve qualified and structured data. We also demonstrate the applicability of the formulated approach by using an industrial case study in manufacturing. Data sets from the manufacturing industry are analyzed to identify data quality problems and detect interesting subsets for hidden information. With the approach formulated, it is possible to get the useful and diagnostic information in a systematic way about component/machine behavior as the basis for decision support and prognostic model development in predictive maintenance.

A Data-Driven Prediction Method for an Early Warning of Coccidiosis in Intensive Livestock Systems: A Preliminary Study.

Simple SummaryThe development of new methods, able to promptly detect the onset of the infection, is highly important to control parasitic infections of the intestinal tract (coccidiosis) in poultry. The early detection of this disease would reduce the use of anticoccidial and antimicrobials drugs, thus lowering the risk of antibiotic resistance. A data-driven machine learning algorithm was built to relate air quality data to the time of enteric disorders. The results show that this procedure has great potential to be used as a rapid technique to detect coccidiosis.Coccidiosis is still one of the major parasitic infections in poultry. It is caused by protozoa of the genus Eimeria, which cause concrete economic losses due to malabsorption, bad feed conversion rate, reduced weight gain, and increased mortality. The greatest damage is registered in commercial poultry farms because birds are reared together in large numbers and high densities. Unfortunately, these enteric pathologies are not preventable, and their diagnosis is only available when the disease is full-blown. For these reasons, the preventive use of anticoccidials—some of these with antimicrobial action—is a common practice in intensive farming, and this type of management leads to the release of drugs in the environment which contributes to the phenomenon of antibiotic resistance. Due to the high relevance of this issue, the early detection of any health problem is of great importance to improve animal welfare in intensive farming. Three prototypes, previously calibrated and adjusted, were developed and tested in three different experimental poultry farms in order to evaluate whether the system was able to identify the coccidia infection in intensive poultry farms early. For this purpose, a data-driven machine learning algorithm was built, and specific critical values of volatile organic compounds (VOCs) were found to be associated with abnormal levels of oocystis count at an early stage of the disease. This result supports the feasibility of building an automatic data-driven machine learning algorithm for an early warning of coccidiosis.

A Systematic Review of Estimating Breast Cancer Recurrence at the Population Level With Administrative Data.

Exact numbers of breast cancer recurrences are currently unknown at the population level, because they are challenging to actively collect. Previously, real-world data such as administrative claims have been used within expert- or data-driven (machine learning) algorithms for estimating cancer recurrence. We present the first systematic review and meta-analysis, to our knowledge, of publications estimating breast cancer recurrence at the population level using algorithms based on administrative data. The systematic literature search followed Preferred Reporting Items for Systematic Reviews and Meta-Analysis guidelines. We evaluated and compared sensitivity, specificity, positive predictive value, negative predictive value, and overall accuracy of algorithms. A random-effects meta-analysis was performed using a generalized linear mixed model to obtain a pooled estimate of accuracy. Seventeen articles met the inclusion criteria. Most articles used information from medical files as the gold standard, defined as any recurrence. Two studies included bone metastases only in the definition of recurrence. Fewer studies used a model-based approach (decision trees or logistic regression) (41.2%) compared with studies using detection rules without specified model (58.8%). The generalized linear mixed model for all recurrence types reported an accuracy of 92.2% (95% confidence interval = 88.4% to 94.8%). Publications reporting algorithms for detecting breast cancer recurrence are limited in number and heterogeneous. A thorough analysis of the existing algorithms demonstrated the need for more standardization and validation. The meta-analysis reported a high accuracy overall, which indicates algorithms as promising tools to identify breast cancer recurrence at the population level. The rule-based approach combined with emerging machine learning algorithms could be interesting to explore in the future.

Improved shift-invariant sparse coding for noise attenuation of magnetotelluric data

Magnetotelluric (MT) method is widely used for revealing deep electrical structure. However, natural MT signals are susceptible to cultural noises. In particular, the existing data-processing methods usually fail to work when MT data are contaminated by persistent or coherent noises. To improve the quality of MT data collected with strong ambient noises, we propose a novel time-series editing method based on the improved shift-invariant sparse coding (ISISC), a data-driven machine learning algorithm. First, a redundant dictionary is learned autonomously from the raw MT data. Second, cultural noises are reconstructed using the learned dictionary and the orthogonal matching pursuit (OMP) algorithm. Finally, the de-noised MT data are obtained by subtracting the reconstructed cultural noises from the raw MT data. The synthetic data, field experimental data and measured data are tested to verify the effectiveness of the newly proposed method. The results show that our new scheme can effectively remove strong cultural noises and has better adaptability and efficiency than the predefined dictionary-based methods. The method can be used as an alternative when a remote reference station is not available.