Conventional Machine Learning Research Articles

Machine learning models for abstract screening task - A systematic literature review application for health economics and outcome research

ObjectiveSystematic literature reviews (SLRs) are critical for life-science research. However, the manual selection and retrieval of relevant publications can be a time-consuming process. This study aims to (1) develop two disease-specific annotated corpora, one for human papillomavirus (HPV) associated diseases and the other for pneumococcal-associated pediatric diseases (PAPD), and (2) optimize machine- and deep-learning models to facilitate automation of the SLR abstract screening.MethodsThis study constructed two disease-specific SLR screening corpora for HPV and PAPD, which contained citation metadata and corresponding abstracts. Performance was evaluated using precision, recall, accuracy, and F1-score of multiple combinations of machine- and deep-learning algorithms and features such as keywords and MeSH terms.Results and conclusionsThe HPV corpus contained 1697 entries, with 538 relevant and 1159 irrelevant articles. The PAPD corpus included 2865 entries, with 711 relevant and 2154 irrelevant articles. Adding additional features beyond title and abstract improved the performance (measured in Accuracy) of machine learning models by 3% for HPV corpus and 2% for PAPD corpus. Transformer-based deep learning models that consistently outperformed conventional machine learning algorithms, highlighting the strength of domain-specific pre-trained language models for SLR abstract screening. This study provides a foundation for the development of more intelligent SLR systems.

Detecting emotions through EEG signals based on modified convolutional fuzzy neural network

Emotion is a human sense that can influence an individual’s life quality in both positive and negative ways. The ability to distinguish different types of emotion can lead researchers to estimate the current situation of patients or the probability of future disease. Recognizing emotions from images have problems concealing their feeling by modifying their facial expressions. This led researchers to consider Electroencephalography (EEG) signals for more accurate emotion detection. However, the complexity of EEG recordings and data analysis using conventional machine learning algorithms caused inconsistent emotion recognition. Therefore, utilizing hybrid deep learning models and other techniques has become common due to their ability to analyze complicated data and achieve higher performance by integrating diverse features of the models. However, researchers prioritize models with fewer parameters to achieve the highest average accuracy. This study improves the Convolutional Fuzzy Neural Network (CFNN) for emotion recognition using EEG signals to achieve a reliable detection system. Initially, the pre-processing and feature extraction phases are implemented to obtain noiseless and informative data. Then, the CFNN with modified architecture is trained to classify emotions. Several parametric and comparative experiments are performed. The proposed model achieved reliable performance for emotion recognition with an average accuracy of 98.21% and 98.08% for valence (pleasantness) and arousal (intensity), respectively, and outperformed state-of-the-art methods.

Sampling-based novel heterogeneous multi-layer stacking ensemble method for telecom customer churn prediction

In recent times, customer churn has become one of the most significant issues in business-oriented sectors with telecommunication being no exception. Maintaining current customers is particularly valuable due to the high degree of rivalry among telecommunication companies and the costs of acquiring new ones. The early prediction of churned customers may help telecommunication companies to identify the causes of churn and design industrial tactics to address or mitigate the churn problem. Controlling customer churn by developing efficient and reliable customer churn prediction (CCP) solutions is essential to achieving this objective. Findings from existing CCP studies have shown that numerous methods, such as rule-based and machine-learning (ML) mechanisms, have been devised to solve the CCP problem. Nonetheless, the problems of adaptability and the resilience of rule-based CCP solutions are its major weaknesses, and the skewed pattern of churn datasets (class imbalance) is detrimental to the prediction performances of conventional ML models in CCP. Hence, this research developed a robust heterogeneous multi-layer stacking ensemble method (HMSE) for effective CCP. Specifically, in the HMSE method, the prediction prowess of five ML classifiers (Random Forest (RF), Bayesian network (BN), Support Vector Machine (SVM), K-Nearest Neighbour (KNN), and Repeated Incremental Pruning to Produce Error Reduction (RIPPER)) with distinct computational characteristics are ensembled based on stacking and the resulting model is further enhanced using a forest penalizing attribute (FPA) model. The synthetic minority oversampling technique (SMOTE) is integrated with the proposed HMSE to balance the skewed class label present in the original experimental datasets. Extensive tests were carried out to determine the effectiveness of the proposed HMSE and S-HMSE on standard telecom CCP datasets. Observed findings from the experimental results showed that HMSE and S-HMSE can effectively predict churners even with the class imbalance (skewed datasets) problem. In addition, comparison studies demonstrated that the suggested S-HMSE offered improved prediction performance and optimum solutions for CCP in the telecom sector in comparison with baseline classifiers, homogeneous ensemble methods, and current CCP approaches.

ELRL-MD: a deep learning approach for myocarditis diagnosis using cardiac magnetic resonance images with ensemble and reinforcement learning integration.

Myocarditis poses a significant health risk, often precipitated by viral infections like Coronavirus disease (COVID-19), and can lead to fatal cardiac complications. As a less invasive alternative to the standard diagnostic practice of endomyocardial biopsy, which is highly invasive and thus limited to severe cases, Cardiac Magnetic Resonance (CMR) imaging offers a promising solution for detecting myocardial abnormalities. This study introduces a deep model called ELRL-MD that combines ensemble learning and Reinforcement Learning (RL) for effective myocarditis diagnosis from CMR images. The model begins with pre-training via the Artificial Bee Colony (ABC) algorithm to enhance the starting point for learning. An array of Convolutional Neural Networks (CNNs) then works in concert to extract and integrate features from CMR images for accurate diagnosis. Leveraging the Z-Alizadeh Sani myocarditis CMR dataset, the model employs RL to navigate the dataset's imbalance by conceptualizing diagnosis as a decision-making process. ELRL-DM demonstrates remarkable efficacy, surpassing other deep learning, conventional machine learning, and transfer learning models, achieving an F-measure of 88.2\% and a geometric mean of 90.6\%. Extensive experimentation helped pinpoint the optimal reward function settings and the perfect count of CNNs. The study addresses the primary technical challenge of inherent data imbalance in CMR imaging datasets and the risk of models converging on local optima due to suboptimal initial weight settings. Further analysis, leaving out ABC and RL components, confirmed their contributions to the model's overall performance, underscoring the effectiveness of addressing these critical technical challenges.

Independent Video Steganalysis Framework Perspective of Secure Video Processing

We have designed an algorithm of cross-domain feature set extraction for cross-domain video steganalysis of video steganography in multiple domains. For video steganography, we implemented the recently proposed steganography methods such as PMs, MVs, and IPMs. The outcomes of these techniques have fed as input to the proposed Steganalysis approach. We have designed a cross-domain technique for Steganalysis in which the global feature set is extracted according to common statistical properties. After the extraction of global features, the domain-specific features have been extracted in the local features set. Both global and local features are set to form the cross-domain steganalysis technique. The classification has been performed by using the conventional machine learning classifiers such as Support Vector Machine (SVM) and Artificial Neural Network (ANN).

Music Genre Classification

The Music Genre Classification model automatically divides music into different genres using a small number of audio files and a range of musical attributes. This topic is highly relevant to the field of music information retrieval since it provides a way to organize and analyze large amounts of music files. For MGC, standard machine learning techniques such as SVM, KNN, Decision trees, and neural networks can be applied. These algorithms are trained to recognize different musical qualities and traits, which allows them to categorize the audio files into different genres. Numerous applications show that deep learning algorithms—such as CNN, ANN, and others—perform better than conventional machine learning algorithms. Consequently, the CNN method is adjusted to perform the categorization of music files. This classifies musical genres using deep learning methods from CNN. To evaluate the effectiveness of the MGC algorithms, accuracy is used. Moreover, the impact of different algorithms on MGC performance can be compared and studied. It can be applied to automated music recommendation systems, music production, and music education.

Machine learning-based characterization of friction stir welding in aluminum alloys

This research presents a machine learning (ML) approach that integrates Bayesian theorem to predict residual stress, plastic deformation, and peak temperature in friction stir welding (FSW) of different types of aluminum alloys. The training data was acquired from extensive numerical simulations of FSW conducted across various input parameter configurations. The regression analysis exhibits strong predictive performance for residual stress, peak temperature, and plastic deformation, achieving determination coefficients of 0.969, 0.955, and 0.919, respectively. The high prediction performance of the developed model is primarily attributed to Bayesian optimization, distinguishing it from other conventional ML methods. The results further demonstrate the efficacy of the proposed model in assessing the relevance of input features to prediction efficiency and output targets. For instance, concerning processing parameters, the predictive performance is closely correlated with the significance of preheating, traverse speed, and rotational speed in FSW, compared to welding force and plate thickness. On the other hand, considering material properties, it was observed that the hardness value of aluminum alloys emerged as a crucial predictor in the model so that alloys with moderate hardness values showed greater welding consistency and significantly enhanced prediction accuracy.

Cortical signals analysis to recognize intralimb mobility using modified RNN and various EEG quantities

Electroencephalogram (EEG) signals are critical in interpreting sensorimotor activities for predicting body movements. However, their efficacy in identifying intralimb movements, such as the dorsiflexion and plantar flexion of the foot, remains suboptimal. This study aims to explore whether various EEG signal quantities can effectively recognize intralimb movements to facilitate the development of Brain-Computer Interface (BCI) devices for foot rehabilitation. This research involved twenty-two healthy, right-handed participants. EEG data were collected using 21 electrodes positioned over the motor cortex, while two electromyography (EMG) electrodes recorded the onset of ankle joint movements. The study focused on analyzing slow cortical potential (SCP) and sensorimotor rhythms (SMR) in alpha and beta bands from the EEG. Five key features—fourth-order Autoregressive feature, variance, waveform length, standard deviation, and permutation entropy—were extracted. A modified Recurrent Neural Network (RNN) including Long Short-term Memory (LSTM) and Gated Recurrent Unit (GRU) algorithms was developed for movement recognition. These were compared against conventional machine learning algorithms, including nonlinear Support Vector Machine (SVM) and k Nearest Neighbourhood (kNN) classifiers. The performance of the proposed models was assessed using two data schemes: within-subject and across-subjects. The findings demonstrated that the GRU and LSTM models significantly outperformed traditional machine learning algorithms in recognizing different EEG signal quantities for intralimb movement. The study indicates that deep learning models, particularly GRU and LSTM, hold superior potential over standard machine learning techniques in identifying intralimb movements using EEG signals. Where the accuracies of LSTM for within and across subjects were 98.87 ± 1.80 % and 87.38 ± 0.86 % respectively. Whereas the accuracy of GRU within and across subjects were 99.18 ± 1.28 % and 86.44 ± 0.69 % respectively. This advancement could significantly benefit the development of BCI devices aimed at foot rehabilitation, suggesting a new avenue for enhancing physical therapy outcomes.

A Novel Quantum Neural Network Approach to Combating Fake Reviews

AbstractAs e-commerce has grown gradually online item assessments have emerged as a key source of consumer data. That stated, there are problems with the consistency and fictitiousness of the evaluations because there are numerous fake or fraudulent ones. These misleading assessments are generated during the investigation in an attempt to mislead customers about the nature of a real advantage, compromising their ability to make a predetermined decision and damaging the reputations of businesses. A cutting-edge interrogation department revealed that quantum machine learning (QML) could manage a huge amount of machine-trained data and could convey almost emotional choices in the context of inaccurate checks. It is truly beneficial in obtaining reviews for things that are incorrect. Opinion, generating trends, interpersonal relationships, and assessing fatigue is merely a few of the attributes. Tests conducted utilizing the Amazon fraudulent review. The dataset demonstrates that QML tactics outperform conventional knowledge acquisition procedures in errands, including the place of fraudulent reviews. The delicacy and tolerance of incorrect review distinguishing evidence can be significantly advanced, although QML is still in its early stages of development. Both our proposed model and model pass rigorous conventional machine learning algorithms testing with a remarkable level of accuracy. An article introduces a unique approach to fraudulent review detection and brings in the QNN algorithm as a solution. A deep learning model, such as this one, has an 86% accuracy rate in quantum computer implementation, which is an impressive level of innovation, especially if it comes with successful results. Involvement in these cutting-edge technologies promises significant benefits in battling the problem of false identities on the Web. In our case, the experimental results demonstrate that our QNN algorithm, which can accurately identify fake reviews, will become a key weapon for suppressing various forms of fraudulence on emerging digital technology platforms.

Advanced AI-driven approach for enhanced brain tumor detection from MRI images utilizing EfficientNetB2 with equalization and homomorphic filtering

Brain tumors pose a significant medical challenge necessitating precise detection and diagnosis, especially in Magnetic resonance imaging(MRI). Current methodologies reliant on traditional image processing and conventional machine learning encounter hurdles in accurately discerning tumor regions within intricate MRI scans, often susceptible to noise and varying image quality. The advent of artificial intelligence (AI) has revolutionized various aspects of healthcare, providing innovative solutions for diagnostics and treatment strategies. This paper introduces a novel AI-driven methodology for brain tumor detection from MRI images, leveraging the EfficientNetB2 deep learning architecture. Our approach incorporates advanced image preprocessing techniques, including image cropping, equalization, and the application of homomorphic filters, to enhance the quality of MRI data for more accurate tumor detection. The proposed model exhibits substantial performance enhancement by demonstrating validation accuracies of 99.83%, 99.75%, and 99.2% on BD-BrainTumor, Brain-tumor-detection, and Brain-MRI-images-for-brain-tumor-detection datasets respectively, this research holds promise for refined clinical diagnostics and patient care, fostering more accurate and reliable brain tumor identification from MRI images. All data is available on Github: https://github.com/muskan258/Brain-Tumor-Detection-from-MRI-Images-Utilizing-EfficientNetB2).

Topological Deep Learning Model for Thyroid Multi-Class Categorization

This study presents an innovative method for categorizing thyroid disorders using a topological deep learning (TDL) model. The strategy utilizes supervised learning and places particular emphasis on topological data analysis (TDA). The work employs a dataset obtained from the WEKA simulator, consisting of 32 characteristics and 3897 instances. It creates a predictive model using the WEKA 3.8.4 simulator, deviating from conventional machine learning methods by giving priority to topological features to improve prediction and classification. The TDL model includes the steps of data preprocessing, selecting unique features, and using TDL algorithms such as Naive Bayes, Decision Trees, and Artificial Neural Networks. The model utilizes embedding to extract information, investigates the importance of specific dimensions, and applies an algorithm to make accurate predictions. The results demonstrate excellent performance, with accuracy, recall, F1-score, and ROC values of 0.96, 0.97, 0.95, and 0.96, respectively. The multi-class classification study provides additional validation of the model's effectiveness, demonstrating an exceptional accuracy of 97.83% and a minimal error rate of 0.21. This highlights the model's precision and dependability, making it suitable for possible clinical applications. This study highlights the efficacy of using topological characteristics and topological data analysis (TDA) methods in the classification of thyroid conditions into many classes. This approach shows potential for enhancing accuracy and dependability in intricate classification tasks.

Interpretable machine learning of SPAC system via a mechanism-assisted gaussian process group: Representation of the system mechanism by data

Traditional models of the soil-plant-atmosphere continuum (SPAC) system are physics-based and their practical application has been hindered by issues of high parameterization, prior structural bias, and costly running. In this paper, we developed a machine learning modelling method based on the Gaussian process (GP) to avoid these difficulties of traditional modelling. At the same time, shortcomings of conventional machine learning in terms of interpretability and mechanism representation were addressed. The innovative method consisted of structuring a local GP group framework and improving kernel functions used in each local GP. The resultant model was endowed with advanced capacities, including representing interplays between subprocesses within the complex SPAC system, representing nonstationary subprocess dynamics caused by crop growth stage shifts, as well as automatically interpreting key low-order input interactions and dominant input variables for each subprocess. The performance of our model was examined on synthetic SPAC system datasets that covered three different soil conditions. Results demonstrated that its interpretations regarding subprocess mechanisms were robust across different soil conditions and consistent with domain knowledge. Compared to a conventional global GP model and two deep learning models (DNN and LSTM), our model performed significantly better not only in regular prediction experiments but also in generalization experiments that required models to be transferred across different conditions. Our study suggested that it is promising to enable a machine learning model interpretable by improving its feature representation function. For machine learning modelling of complex systems like the SPAC, it may be more important to enhance the model’s mechanism representation capability than simply using deeper networks.

An Improved Network Intrusion Detection Method Based On CNN-LSTM-SA

Network intrusion detection is an essential component of contemporary cybersecurity strategies, and the development of efficient techniques to accurately identify malicious activities has become a priority. This study investigates the performance of various conventional machine learning algorithms, including decision trees, naive Bayes, naive Bayes trees, random forest, random trees, MLP, and SVM, in detecting network intrusions using binary and multi-classification approaches. Furthermore, the study proposes a deep learning method, CNN-LSTM-SA, which consistently outperforms conventional machine learning techniques in terms of precision, recall, F1 score, and overall accuracy for network intrusion detection. Specifically, the proposed method combines CNN and LSTM with SA in machine learning theory to extract more optimized, strongly correlated features. The proposed method is evaluated using the benchmark NSL-KDD database. The results indicate that the CNN-LSTM-SA method holds great potential in enhancing the efficacy of network intrusion detection systems.

Combining LiDAR and Spaceborne Multispectral Data for Mapping Successional Forest Stages in Subtropical Forests

The Brazilian Atlantic Rainforest presents great diversity of flora and stand structures, making it difficult for traditional forest inventories to collect reliable and recurrent information to classify forest succession stages. In recent years, remote sensing data have been explored to save time and effort in classifying successional forest stages. However, there is a need to understand if any of these sensors stand out for this purpose. Here, we evaluate the use of multispectral satellite data from four different platforms (CBERS-4A, Landsat-8/OLI, PlanetScope, and Sentinel-2) and airborne light detection and ranging (LiDAR) to classify three forest succession stages in a subtropical ombrophilous mixed forest located in southern Brazil. Different features extracted from multispectral and LiDAR data, such as spectral bands, vegetation indices, texture features, and the canopy height model (CHM) and LiDAR intensity, were explored using two conventional machine learning methods such as random trees (RT) and support vector machine (SVM). The statistically based maximum likelihood (MLC) algorithm was also compared. The classification accuracy was evaluated by generating a confusion matrix and calculating the kappa index and standard deviation based on field measurements and unmanned aerial vehicle (UAV) data. Our results show that the kappa index ranged from 0.48 to 0.95, depending on the chosen dataset and method. The best result was obtained using the SVM algorithm associated with spectral bands, CHM, LiDAR intensity, and vegetation indices, regardless of the sensor. Datasets with Landsat-8 or Sentinel-2 information performed better results than other optical sensors, which may be due to the higher intraclass variability and less spectral bands in CBERS-4A and PlanetScope data. We found that the height information derived from airborne LiDAR and its intensity combined with the multispectral data increased the classification accuracy. However, the results were also satisfactory when using only multispectral data. These results highlight the potential of using freely available satellite information and open-source software to optimize forest inventories and monitoring, enabling a better understanding of forest structure and potentially supporting forest management initiatives and environmental licensing programs.

Low-cost and efficient prediction hardware for tabular data using tiny classifier circuits

AbstractA typical machine learning development cycle maximizes performance during model training and then minimizes the memory and area footprint of the trained model for deployment on processing cores, graphics processing units, microcontrollers or custom hardware accelerators. However, this becomes increasingly difficult as machine learning models grow larger and more complex. Here we report a methodology for automatically generating predictor circuits for the classification of tabular data. The approach offers comparable prediction performance to conventional machine learning techniques as substantially fewer hardware resources and power are used. We use an evolutionary algorithm to search over the space of logic gates and automatically generate a classifier circuit with maximized training prediction accuracy, which consists of no more than 300 logic gates. When simulated as a silicon chip, our tiny classifiers use 8–18 times less area and 4–8 times less power than the best-performing machine learning baseline. When implemented as a low-cost chip on a flexible substrate, they occupy 10–75 times less area, consume 13–75 times less power and have 6 times better yield than the most hardware-efficient ML baseline.

Enhancing photovoltaic energy forecasting: a progressive approach using wavelet packet decomposition

Abstract Accurate photovoltaic (PV) energy forecasting plays a crucial role in the efficient operation of PV power stations. This study presents a novel hybrid machine-learning (ML) model that combines Gaussian process regression with wavelet packet decomposition to forecast PV power half an hour ahead. The proposed technique was applied to the PV energy database of a station located in Algeria and its performance was compared to that of traditional forecasting models. Performance evaluations demonstrate the superiority of the proposed approach over conventional ML methods, including Gaussian process regression, extreme learning machines, artificial neural networks and support vector machines, across all seasons. The proposed model exhibits lower normalized root mean square error (nRMSE) (2.116%) and root mean square error (RMSE) (208.233 kW) values, along with a higher coefficient of determination (R2) of 99.881%. Furthermore, the exceptional performance of the model is maintained even when tested with various prediction horizons. However, as the forecast horizon extends from 1.5 to 5.5 hours, the prediction accuracy decreases, evident by the increase in the RMSE (710.839 kW) and nRMSE (7.276%), and a decrease in R2 (98.462%). Comparative analysis with recent studies reveals that our approach consistently delivers competitive or superior results. This study provides empirical evidence supporting the effectiveness of the proposed hybrid ML model, suggesting its potential as a reliable tool for enhancing PV power forecasting accuracy, thereby contributing to more efficient grid management.

Pore Pressure Prediction for High-Pressure Tight Sandstone in the Huizhou Sag, Pearl River Mouth Basin, China: A Machine Learning-Based Approach

A growing number of large data sets have created challenges for the oil and gas industry in predicting reservoir parameters and assessing well productivity through efficient and cost-effective techniques. The design of drilling plans for a high-pressure tight-sand reservoir requires accurate estimations of pore pressure (Pp) and reservoir parameters. The objective of this study is to predict and compare the Pp of Huizhou Sag, Pearl River Mouth Basin, China, using conventional techniques and machine learning (ML) algorithms. We investigated the characteristics of low-permeability reservoirs by observing well-logging data sets and cores and examining thin sections under a microscope. In the reservoir zone, the average hydrocarbon saturation is 55%, and the average effective porosity is 11%. The tight sandstone reservoirs consist of fine- to extremely fine-grained argillaceous feldspathic sandstone. The mean absolute error for reservoir property prediction is 1.3%, 2.2%, and 4.8%, respectively, for effective porosity, shale volume, and water saturation. Moreover, the ML algorithm was employed to cross-check the validity of the prediction of Pp. Combining conventional and ML techniques with the core data demonstrates a correlation coefficient (R2) of 0.9587, indicating that ML techniques are the most effective in testing well data. This study shows that ML can effectively predict Pp at subsequent depths in adjacent geologically similar locations. Compared to conventional methods, a substantial data set and ML algorithms improve the precision of Pp predictions.

Enhancing robotic manipulator fault detection with advanced machine learning techniques

The optimization of rotating machinery processes is crucial for enhanced industrial productivity. Automatic machine health monitoring systems play a vital role in ensuring smooth operations. This study introduces a novel approach for fault diagnosis in robotic manipulators through motor sound analysis to enhance industrial efficiency and prevent machinery downtime. A unique dataset is generated using a custom robotic manipulator to examine the effectiveness of both deep learning and traditional machine learning in identifying motor anomalies. The investigation includes a two-stage analysis, initially leveraging 2D spectrogram features with neural network architectures, followed by an evaluation of 1D MFCC features using various conventional machine learning algorithms. The results reveal that the proposed custom CNN and 1D-CNN models significantly surpass traditional methods, achieving an F1-score exceeding 92%, highlighting the potential of sound analysis for automated fault detection in robotic systems. Additional experiments were carried out to investigate 1D MFCC features with various machine learning algorithms, including KNN, DT, LR, RF, SVM, MLP, and 1D-CNN. Augmented with additional data collected from the locally designed manipulator, our experimental setup significantly enhances model performance. Particularly, the 1D-CNN stands out as the top-performing model on the augmented dataset.

NOx Emission Predictions in Gas Turbines Through Integrated Data-Driven Machine Learning Approaches

Abstract The reduction of NOx emissions is a paramount endeavor in contemporary engineering and energy production, as these emissions are closely linked to adverse environmental and health impacts. The prediction of NOx emission from gas turbines through several integrated data-driven machine learning methods has been evaluated in study. The study compares the performance of ensemble and conventional machine learning models, demonstrating superior accuracy achieved by the ensemble models. Specifically, the Random Forest model achieved an accuracy rate of 91.68%, XGBoost yielded an accuracy of 91.54%, and CATBoost exhibited the highest accuracy at 92.76%. These findings highlight the capability of data-driven machine learning techniques in enhancing NOx emission predictions in gas turbines. The improved prediction by ensembles can be utilized in the development and implementation of more effective control and mitigation strategies in practical applications. Through the application of these advanced machine learning approaches, the gas turbine industry can play a pivotal role in minimizing its environmental impact while optimizing operational efficiency. This study also provides valuable insights into the effectiveness of ensemble machine learning models, advancing our understanding of their capabilities in addressing the critical issue of NOx emissions from gas turbines.

Interpretable Trend Analysis Neural Networks for Longitudinal Data Analysis

Cohort study is one of the most commonly used study methods in medical and public health researches, which result in longitudinal data. Conventional statistical models and machine learning methods are not capable of modeling the evolution trend of the variables in longitudinal data. In this article, we propose a Trend Analysis Neural Networks (TANN), which models the evolution trend of the variables by adaptive feature learning. TANN was tested on dataset of Kaiuan research. The task was to predict occurrence of cardiovascular events within 2 and 5 years, with three repeated medical examinations during 2008 and 2013. For 2-year prediction, The AUC of the TANN is 0.7378, which is a significant improvement than that of conventional methods, while that of TRNS, RNN, DNN, GBDT, RF, and LR are 0.7222, 0.7034, 0.7054, 0.7136, 0.7160, and 0.7024, respectively. For 5-year prediction, TANN also shows improvement. The experimental results show that the proposed TANN achieves better prediction performance on cardiovascular events prediction than conventional models. Furthermore, by analyzing the weights of TANN, we could find out important trends of the indicators, which are ignored by conventional machine learning models. The trend discovery mechanism interprets the model well. TANN is an appropriate balance between high performance and interpretability.