Mutual Information Research Articles

Evaluating Greeting Models for Human-Robot Engagement During First Encounters

Abstract Engagement during first encounters needs both parties to comprehend each other’s intent without mutual information. A robot may fail to engage someone because people are unaware of the robot’s presence or intent to interact, or because they are uninterested in interacting with robots. In this paper, we compared three different greeting strategies for mobile social robots in first encounters with humans using the Kendon’s greeting protocol, which was derived from observations of human greeting behaviors. The theoretical protocol defines a sequence of six phases during greetings: initiation of the approach, distance salutation, head dip, approach, final approach, and close salutation. The three greeting strategies we compare in our empirical studies specifically implement the first three phases of the Kendon’s model (M1), the six phases (M2), or the six phases as a response of the human greeting phase (M3). We performed a user study during a public art exhibition in two different situations: in-the-wild and with invited participants. The results of these two experiments combined both quantitative and qualitative measures and support that people’s engagement and impressions of the robot can be enhanced when the robot employs M2 and M3, compared to M1. Our results generally support that Kendon’s greeting model is effective in human-robot first encounters for mobile social robots. We also discuss important considerations for running in-the-wild robotic experiments.

Fermionic entanglement in the presence of background electric and magnetic fields

In this study, we investigate the fermionic Schwinger effect in the presence of a constant magnetic field within (1+3)-dimensional Minkowski spacetime, considering both constant and pulsed electric fields. We analyze the correlations between Schwinger pairs for the vacuum and maximally entangled states of two fermionic fields. The correlations are quantified using entanglement entropy and Bell’s inequality violation for the vacuum state, while Bell’s inequality violation and mutual information are used for the maximally entangled state. One can observe the variation of the entanglement produced for fermionic modes with respect to different parameters. Additionally, we discuss the key differences from the behaviour of scalar fields in this context. This study offers deeper insights into quantum field theory and the dynamics of entanglement in the fermionic Schwinger effect.

Feature selection for high-dimensional data using a multivariate search space reduction strategy based scatter search

In feature selection, the increasing of the dimensionality and the complexity of feature interactions make the problem challenging. Furthermore, searching for an optimal subset of features from a high-dimensional feature space is known to be an NP-hard problem. To improve the efficiency and effectiveness of the search algorithm, feature grouping has emerged as a way to reduce the search space by clustering features according to a measure. In this work we propose to reduce the search space by applying a greedy algorithm, called Multivariate Greedy Predominant Groups Generator (MGPGG). MGPGG extends the idea of the Greedy Predominant Groups Generator (GPGG) algorithm by taking into account feature interaction among three or more features. For this purpose, MGPGG uses the Multivariate Symmetrical Uncertainty (MSU) to group features that share information about the class label. We also propose a Scatter Search strategy that integrates MGPGG to find small subsets of features with high predictive power. The proposed algorithm, called Multivariate Predominant Group-based Scatter Search (MPGSS), is tested on high-dimensional data from biomedical and text-mining fields. The proposal is compared with state-of-the-art feature selection strategies. Results show that MPGSS is competitive since it is capable of finding small subsets of features while keeping high predictive classification models.

Causal Discovery and Reasoning for Continuous Variables with an Improved Bayesian Network Constructed by Locality Sensitive Hashing and Kernel Density Estimation

The structure learning of a Bayesian network (BN) is a crucial process that aims to unravel the complex dependencies relationships among variables using a given dataset. This paper proposes a new BN structure learning method for data with continuous attribute values. As a non-parametric distribution-free method, kernel density estimation (KDE) is applied in the conditional independence (CI) test. The skeleton of the BN is constructed utilizing the test based on mutual information and conditional mutual information, delineating potential relational connections between parents and children without imposing any distributional assumptions. In the searching stage of BN structure learning, the causal relationships between variables are achieved by using the conditional entropy scoring function and hill-climbing strategy. To further enhance the computational efficiency of our method, we incorporate a locality sensitive hashing (LSH) function into the KDE process. The method speeds up the calculations of KDE while maintaining the precision of the estimates, leading to a notable decrease in the time required for computing mutual information, conditional mutual information, and conditional entropy. A BN classifier (BNC) is established by using the computationally efficient BN learning method. Our experiments demonstrated that KDE using LSH has greatly improved the speed compared to traditional KDE without losing fitting accuracy. This achievement underscores the effectiveness of our method in balancing speed and accuracy. By giving the benchmark networks, the network structure learning accuracy with the proposed method is superior to other traditional structure learning methods. The BNC also demonstrates better accuracy with stronger interpretability compared to conventional classifiers on public datasets.

Enhancing Image Quality in Crop Disease Diagnosis: A Comparative Evaluation of Laplacian and Average Filtering with Established Techniques

The agricultural sector is vital to ensuring global food security, yet plant diseases caused by various environmental factors lead to significant reductions in crop yields. Early detection of such diseases is critical, as crop health directly affects both yield and quality. This research introduces a novel pre-processing algorithm designed to enhance the accuracy of crop disease detection from leaf images. The proposed algorithm is compared against three widely used filtering techniques: Median filtering, Wiener filtering, and Gaussian filtering. Before implementing the proposed approach, a detailed assessment of these conventional methods is conducted.The algorithm integrates Laplacian filtering and average filtering to optimize image pre-processing. The effectiveness of this approach is evaluated using performance metrics such as Peak Signal-to-Noise Ratio (PSNR), Mean Squared Error (MSE), and Mutual Information (MI). Results indicate that the proposed method achieves a 0.34% increase in PSNR, a 0.92% reduction in MSE, and a 0.48% improvement in MI compared to existing techniques. These improvements underscore the algorithm's ability to enhance image quality and outperform traditional methods.The findings suggest that incorporating Laplacian and average filtering into the pre-processing pipeline introduces a more efficient methodology for image enhancement. This approach holds potential for advancing image analysis and interpretation in agriculture and other fields, providing a foundation for further innovation in image processing technologies.

Fusion of EEG and EMG signals for detecting pre-movement intention of sitting and standing in healthy individuals and patients with spinal cord injury

IntroductionRehabilitation devices assist individuals with movement disorders by supporting daily activities and facilitating effective rehabilitation training. Accurate and early motor intention detection is vital for real-time device applications. However, traditional methods of motor intention detection often rely on single-mode signals, such as EEG or EMG alone, which can be limited by low signal quality and reduced stability. This study proposes a multimodal fusion method based on EEG–EMG functional connectivity to detect sitting and standing intentions before movement execution, enabling timely intervention and reducing latency in rehabilitation devices.MethodsEight healthy subjects and five spinal cord injury (SCI) patients performed cue-based sit-to-stand and stand-to-sit transition tasks while EEG and EMG data were recorded simultaneously. We constructed EEG–EMG functional connectivity networks using data epochs from the 1.5-s period prior to movement onset. Pairwise spatial filters were then designed to extract discriminative spatial network topologies. Each filter paired with a support vector machine classifier to classify future movements into one of three classes: sit-to-stand, stand-to-sit, or rest. The final prediction was determined using a majority voting scheme.ResultsAmong the three functional connectivity methods investigated—coherence, Pearson correlation coefficient and mutual information (MI)—the MI-based EEG–EMG network showed the highest decoding performance (94.33%), outperforming both EEG (73.89%) and EMG (89.16%). The robustness of the fusion method was further validated through a fatigue training experiment with healthy subjects. The fusion method achieved 92.87% accuracy during the post-fatigue stage, with no significant difference compared to the pre-fatigue stage (p > 0.05). Additionally, the proposed method using pre-movement windows achieved accuracy comparable to trans-movement windows (p > 0.05 for both pre- and post-fatigue stages). For the SCI patients, the fusion method showed improved accuracy, achieving 87.54% compared to single- modality methods (EEG: 83.03%, EMG: 84.13%), suggesting that the fusion method could be promising for practical rehabilitation applications.ConclusionOur results demonstrated that the proposed multimodal fusion method significantly enhances the performance of detecting human motor intentions. By enabling early detection of sitting and standing intentions, this method holds the potential to offer more accurate and timely interventions within rehabilitation systems.

Improving Lithium-Ion Battery State of Health Estimation with an Integrated Convolutional Neural Network, Gated Recurrent Unit, and Squeeze-and-Excitation Model

Abstract Estimating the State of Health (SOH) of lithium batteries is vital for the safe management of new energy systems. This study leverages voltage, current, and temperature data from the NASA battery dataset to extract health features. The relationship between these features and battery capacity is evaluated using mutual information analysis. An integrated Convolutional Neural Network (CNN), Gated Recurrent Unit (GRU), and Squeeze-and-Excitation (SE) model (CNN-GRU-SE) for estimating battery SOH is proposed. The CNN module identifies local features from the input data, the SE module emphasizes key features, and the GRU module captures temporal dependencies, effectively tracking the battery's health trend over time. The estimation results indicate that the CNN-GRU-SE model reduces the Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and Root Mean Square Error (RMSE) by approximately 3% to 10% compared to the CNN, GRU, and CNN-GRU models. These results confirm the superior estimation capability of the integrated CNN-GRU-SE model. Furthermore, the study underscores the effective integration of the strengths of CNN, SE modules, and GRU, demonstrating its potential application in battery health management.

Large Language Model Approach for Zero-Shot Information Extraction and Clustering of Japanese Radiology Reports: Algorithm Development and Validation.

The application of natural language processing in medicine has increased significantly, including tasks such as information extraction and classification. Natural language processing plays a crucial role in structuring free-form radiology reports, facilitating the interpretation of textual content, and enhancing data utility through clustering techniques. Clustering allows for the identification of similar lesions and disease patterns across a broad dataset, making it useful for aggregating information and discovering new insights in medical imaging. However, most publicly available medical datasets are in English, with limited resources in other languages. This scarcity poses a challenge for development of models geared toward non-English downstream tasks. This study aimed to develop and evaluate an algorithm that uses large language models (LLMs) to extract information from Japanese lung cancer radiology reports and perform clustering analysis. The effectiveness of this approach was assessed and compared with previous supervised methods. This study employed the MedTxt-RR dataset, comprising 135 Japanese radiology reports from 9 radiologists who interpreted the computed tomography images of 15 lung cancer patients obtained from Radiopaedia. Previously used in the NTCIR-16 (NII Testbeds and Community for Information Access Research) shared task for clustering performance competition, this dataset was ideal for comparing the clustering ability of our algorithm with those of previous methods. The dataset was split into 8 cases for development and 7 for testing, respectively. The study's approach involved using the LLM to extract information pertinent to lung cancer findings and transforming it into numeric features for clustering, using the K-means method. Performance was evaluated using 135 reports for information extraction accuracy and 63 test reports for clustering performance. This study focused on the accuracy of automated systems for extracting tumor size, location, and laterality from clinical reports. The clustering performance was evaluated using normalized mutual information, adjusted mutual information , and the Fowlkes-Mallows index for both the development and test data. The tumor size was accurately identified in 99 out of 135 reports (73.3%), with errors in 36 reports (26.7%), primarily due to missing or incorrect size information. Tumor location and laterality were identified with greater accuracy in 112 out of 135 reports (83%); however, 23 reports (17%) contained errors mainly due to empty values or incorrect data. Clustering performance of the test data yielded an normalized mutual information of 0.6414, adjusted mutual information of 0.5598, and Fowlkes-Mallows index of 0.5354. The proposed method demonstrated superior performance across all evaluation metrics compared to previous methods. The unsupervised LLM approach surpassed the existing supervised methods in clustering Japanese radiology reports. These findings suggest that LLMs hold promise for extracting information from radiology reports and integrating it into disease-specific knowledge structures.

Improving Soil Heavy Metal Lead Inversion Through Combined Band Selection Methods: A Case Study in Gejiu City, China

Hyperspectral technology has become increasingly important in monitoring soil heavy metal pollution, yet hyperspectral data often contain substantial band redundancy, and band selection methods are typically limited to single algorithms or simple combinations. Multi-algorithm combinations for band selection remain underutilized. To address this gap, this study, conducted in Gejiu, Yunnan Province, China, proposes a multi-algorithm band selection method to enable the rapid prediction of lead (Pb) contamination levels in soil. To construct a preliminary Pb content prediction model, the initial selection of spectral bands utilized methods including CARS (Competitive Adaptive Reweighted Sampling), GA (Genetic Algorithm), MI (mutual information), SPA (Successive Projections Algorithm), and WOA (Whale Optimization Algorithm). The results indicated that WOA achieved the highest modeling accuracy. Building on this, a combined WOA-based band selection method was developed, including combinations such as WOA-CARS, WOA-GA, WOA-MI, and WOA-SPA, with multi-level band optimization further refined by MI (e.g., WOA-GA-MI, WOA-CARS-MI, WOA-SPA-MI). The results showed that the WOA-GA-MI model exhibited optimal performance, achieving an average R2 of 0.75, with improvements of 0.32, 0.11, and 0.02 over the full-spectrum model, the WOA-selected spectral model, and the WOA-GA model, respectively. Additionally, spectral response analysis identified 22 common bands essential for Pb content inversion. The proposed multi-level combined model not only significantly enhances prediction accuracy but also provides new insights into optimizing hyperspectral band selection, serving as a valuable scientific foundation for assessing soil heavy metal contamination.

Tibetan–Chinese speech-to-speech translation based on discrete units

Speech-to-speech translation (S2ST) has evolved from cascade systems which integrate Automatic Speech Recognition (ASR), Machine Translation (MT), and Text-to-Speech (TTS), to end-to-end models. This evolution has been driven by advancements in model performance and the expansion of cross-lingual speech datasets. Despite the paucity of research on Tibetan speech translation, this paper endeavors to tackle the challenge of Tibetan-to-Chinese direct speech-to-speech translation within the multi-task learning framework, employing self-supervised learning (SSL) and sequence-to-sequence model training. Leveraging HuBERT model to extract discrete units of target speech, we develop a speech-to-unit translation (S2UT) model using an encoder-decoder architecture which subsequently generates speech output through a unit-based vocoder. By employing SSL and utilizing discrete representations as training targets, our approach effectively captures linguistic differences, facilitating direct translation between the two languages. We evaluate the performance of HuBERT model under various configurations to select the optimal setup based on Phone-unit Normalized Mutual Information (PNMI) values. After fine-tuning the chosen HuBERT model on specific corpora, we introduce auxiliary tasks to enhance translation performance. This underscores the pivotal role of multi-task learning in improving overall model efficacy. Experimental results validate the feasibility of Tibetan-to-Chinese S2ST, demonstrating promising translation quality and semantic content preservation, despite limited data availability.

A Resource-Efficient Multi-Entropy Fusion Method and Its Application for EEG-Based Emotion Recognition.

Emotion recognition is an advanced technology for understanding human behavior and psychological states, with extensive applications for mental health monitoring, human-computer interaction, and affective computing. Based on electroencephalography (EEG), the biomedical signals naturally generated by the brain, this work proposes a resource-efficient multi-entropy fusion method for classifying emotional states. First, Discrete Wavelet Transform (DWT) is applied to extract five brain rhythms, i.e., delta, theta, alpha, beta, and gamma, from EEG signals, followed by the acquisition of multi-entropy features, including Spectral Entropy (PSDE), Singular Spectrum Entropy (SSE), Sample Entropy (SE), Fuzzy Entropy (FE), Approximation Entropy (AE), and Permutation Entropy (PE). Then, such entropies are fused into a matrix to represent complex and dynamic characteristics of EEG, denoted as the Brain Rhythm Entropy Matrix (BREM). Next, Dynamic Time Warping (DTW), Mutual Information (MI), the Spearman Correlation Coefficient (SCC), and the Jaccard Similarity Coefficient (JSC) are applied to measure the similarity between the unknown testing BREM data and positive/negative emotional samples for classification. Experiments were conducted using the DEAP dataset, aiming to find a suitable scheme regarding similarity measures, time windows, and input numbers of channel data. The results reveal that DTW yields the best performance in similarity measures with a 5 s window. In addition, the single-channel input mode outperforms the single-region mode. The proposed method achieves 84.62% and 82.48% accuracy in arousal and valence classification tasks, respectively, indicating its effectiveness in reducing data dimensionality and computational complexity while maintaining an accuracy of over 80%. Such performances are remarkable when considering limited data resources as a concern, which opens possibilities for an innovative entropy fusion method that can help to design portable EEG-based emotion-aware devices for daily usage.

A multi-constraint representation learning model for identification of ovarian cancer with missing laboratory indicators.

To evaluate the performance of a multi-constraint representation learning classification model for identifying ovarian cancer with missing laboratory indicators. Tabular data with missing laboratory indicators were collected from 393 patients with ovarian cancer and 1951 control patients. The missing ovarian cancer laboratory indicator features were projected to the latent space to obtain a classification model using the representational learning classification model based on discriminative learning and mutual information coupled with feature projection significance score consistency and missing location estimation. The proposed constraint term was ablated experimentally to assess the feasibility and validity of the constraint term by accuracy, area under the ROC curve (AUC), sensitivity, and specificity. Cross-validation methods and accuracy, AUC, sensitivity and specificity were also used to evaluate the discriminative performance of this classification model in comparison with other interpolation methods for processing of the missing data. The results of the ablation experiments showed good compatibility among the constraints, and each constraint had good robustness. The cross-validation experiment showed that for identification of ovarian cancer with missing laboratory indicators, the AUC, accuracy, sensitivity and specificity of the proposed multi-constraints representation-based learning classification model was 0.915, 0.888, 0.774, and 0.910, respectively, and its AUC and sensitivity were superior to those of other interpolation methods. The proposed model has excellent discriminatory ability with better performance than other missing data interpolation methods for identification of ovarian cancer with missing laboratory indicators.

Quality and reliability of IFC/BIM models for public educational facilities construction projects via clash detection

This study investigates the reliability of monodiscipline IFC/BIM models in public construction projects of educational facilities through advanced clash detection and quantitative analysis. Data were collected from BIM models of two kindergartens and a school in Vilnius, Lithuania, representing different design disciplines. A mixed-methods approach was employed to analyse the number, types, and geometric characteristics of detected clashes. The research introduces innovative metrics, such as the Relative Quality Coefficient (RQC), Relative Uncertainty Coefficient (RUC), and Modified Relative Quality Coefficient (MRQC), to assess model quality and reliability quantitatively. The findings reveal a direct relationship between model complexity, clash detection precision, and the number of identified clashes, underscoring the importance of enhanced quality control measures in IFC/BIM models for public procurement. The study concludes that the implementation of these novel metrics can enhance the reliability of IFC/BIM models, thereby optimizing the design and construction process.

A privacy-enhanced framework with deep learning for botnet detection

A botnet is a group of hijacked devices that conduct various cyberattacks, which is one of the most dangerous threats on the internet. Organizations or individuals use network traffic to mine botnet communication behavior features. Network traffic often contains individual users’ private information, such as website passwords, personally identifiable information, and communication content. Among the existing botnet detection methods, whether they extract deterministic traffic interaction features, use DNS traffic, or methods based on raw traffic bytes, these methods focus on the detection performance of the detection model and ignore possible privacy leaks. And most methods are combined with machine learning and deep learning technologies, which require a large amount of training data to obtain high-precision detection models. Therefore, preventing malicious persons from stealing data to infer privacy during the botnet detection process has become an issue worth pondering. Based on this problem, this article proposes a privacy-enhanced framework with deep learning for botnet detection. The goal of this framework is to learn a feature extractor. It can hide the private information that the attack model tries to infer from the intermediate anonymity features, while maximally retaining the interactive behavior features contained in the original traffic for botnet detection. We design a privacy confrontation algorithm based on a mutual information calculation mechanism. This algorithm simulates the game between the attacker trying to infer private information through the attack model and the data processor retaining the original content of the traffic to the maximum extent. In order to further ensure the privacy protection of the feature extractor during the training process, we train the feature extractor in the federated learning training mode. We extensively evaluate our approach, validating it on two public datasets and comparing it with existing methods. The results show that our method can effectively ensure detection accuracy on the basis of removing private information. For the CTU-13 dataset, the detection framework achieves the best detection performance; for the ISCX-2014 dataset, the accuracy of the framework is less than 1% lower than the best effect.

Development of improved deep learning models for multi-step ahead forecasting of daily river water temperature

Precise river water temperature (WT) forecasts are essential for monitoring water quality. This study addresses the limited use of signal decomposition in hybrid WT prediction models by proposing three methods: namely ensemble empirical mode decomposition (EEMD) on AdaBoost, long short-term memory (LSTM), and gated recurrent unit (GRU). These models integrate ensemble empirical mode decomposition (EEMD) with machine learning techniques for forecasting WT across multiple time horizons (one, three, and five days). The performance of implemented models were tested on two river stations located on the Clackamas River (USGS 14211010) and Willamette River (USGS 14211720). Some error measures comprising root mean square (RMSE), mean absolute error (MAE), coefficient of determination (R2), uncertainty coefficient in 95% confidence level (U95%), and mean absolute percentage error (MAPE) were applied in assessing the models’ performances. The performance of the proposed merged methods, including EEMD-AdaBoost, EEMD-LSTM, and EEMD-GRU, were compared with their simple forms. The outcomes illustrated that the hybrid models performed better than the relevant individual methods; however, the river WT forecasts of EEMD-LSTM and EEMD-GRU were found to be much closer to the observed data than those of the EEMD-AdaBoost method. The better accuracy of hybrid models compared to their corresponding simple ones can be explained by considering the potential of EEMD in separating intrinsic patterns and reducing the noises, leading to reliable forecasts of river WT time series. A performance comparison of the simple models also denoted the superiority of LSTM and GRU over the AdaBoost. The superior river WT forecasts at both stations during the testing stage were concluded for one day ahead at EEMD-GRU model (USGS 14211010: RMSE = 0.1929 ℃, MAE = 0.1489 ℃, R2 = 0.9988, U95% = 0.3745, MAPE = 1.3608%; USGS 14211720: RMSE = 0.1918 ℃, MAE = 0.1558 ℃, R2 = 0.9990, U95% = 0.3690, MAPE = 1.1790%).

Characterization of Spatial Cognitive EEG Signals Using Normalized Adjusted Permutation Conditional Mutual Information

Spatial cognitive ability, a fundamental domain within the human cognitive system, involves the perception, interpretation, and manipulation of spatial environments. This study introduces a new EEG feature extraction algorithm, Normalized Adjusted Permutation Conditional Mutual Information (NAPCMI), to improve the accuracy of spatial cognition assessments. By capturing the symmetry and temporal dependencies within EEG signals during spatial cognition tasks, NAPCMI enhances the ability to extract relevant features. The study validates NAPCMI using a BCI-VR spatial cognition assessment system, incorporating gesture recognition. Results demonstrate that NAPCMI outperforms traditional methods in feature extraction, highlighting its potential for advancing the understanding and assessment of spatial cognitive abilities. The findings also emphasize the significance of specific EEG frequency bands, such as Delta and Beta1, in spatial cognition tasks, further validating NAPCMI’s effectiveness.

Quantum Broadcast Channel Simulation via Multipartite Convex Splitting

We show that the communication cost of quantum broadcast channel simulation under free entanglement assistance between the sender and the receivers is asymptotically characterized by an efficiently computable single-letter formula in terms of the channel’s multipartite mutual information. Our core contribution is a new one-shot achievability result for multipartite quantum state splitting via multipartite convex splitting. As part of this, we face a general instance of the quantum joint typicality problem with arbitrarily overlapping marginals. The crucial technical ingredient to sidestep this difficulty is a conceptually novel multipartite mean-zero decomposition lemma, together with employing recently introduced complex interpolation techniques for sandwiched Rényi divergences. Moreover, we establish an exponential convergence of the simulation error when the communication costs are within the interior of the capacity region. As the costs approach the boundary of the capacity region moderately quickly, we show that the error still vanishes asymptotically.

Quantitative Analysis of Multi-Elements in a Micron-Sized Single Particle Based on Laser-Induced Breakdown Spectroscopy Signal Enhancement of an Optical Fiber Collimated System.

With rapid, energy-intensive, and coal-fueled economic growth, global air quality is deteriorating, and particulate matter pollution has emerged as one of the major public health problems worldwide. It is extremely urgent to achieve carbon emission reduction and air pollution prevention and control, aiming at the common problem of weak and unstable signals of characteristic elements in the application of laser-induced breakdown spectroscopy (LIBS) technology for trace element detection. In this study, the influence of the optical fiber collimation signal enhancement method on the LIBS signal was explored. Then, the influence of the LIBS signal enhancement system based on an optical fiber collimated system on LIBS spectral signal intensity and signal-to-noise ratio (SNR) was compared, and the influences of different spectral preprocessing methods and different variable selection methods on the prediction performance of the random forest (RF) calibration model were investigated. Finally, the Savitzky-Golay convolution derivative (SG)-variable importance projection (VIP)-mutual information (MI)-RF (Zn), first-order derivative (D1st)-variable importance measurement (VIM)-successive projections algorithm (SPA)-RF (Cu), and D1st-VIM-MI-RF (Ni) optimal models were constructed according to the optimal spectral preprocessing method and the optimal hybrid variable selection method. The prediction performances of their optimal RF model after SG-VIP-MI (Zn), D1st-VIM-SPA (Cu), and D1st-VIM-MI (Ni) spectral preprocessing and hybrid variable selection method are presented as follows: Zn (Rp2 = 0.9860; MREP = 0.0590), Cu (Rp2 = 0.9817; MREP = 0.0405), and Ni (Rp2 = 0.9856; MREP = 0.0875). The above results demonstrate that the RF calibration model based on the optical fiber collimated LIBS signal enhancement method, the optimal spectral preprocessing method, and variable selection strategy overcome the key problems of low SNR and low quantitative accuracy in single particle detection. It is expected to provide a theoretical basis and technical support for in situ online rapid monitoring of particulate matter.

Predictive value of a constructed artificial neural network model for microvascular invasion in hepatocellular carcinoma: A retrospective study.

Microvascular invasion (MVI) is a significant risk factor for recurrence and metastasis following hepatocellular carcinoma (HCC) surgery. Currently, there is a paucity of preoperative evaluation approaches for MVI. To investigate the predictive value of texture features and radiological signs based on multiparametric magnetic resonance imaging in the non-invasive preoperative prediction of MVI in HCC. Clinical data from 97 HCC patients were retrospectively collected from January 2019 to July 2022 at our hospital. Patients were classified into two groups: MVI-positive (n = 57) and MVI-negative (n = 40), based on postoperative pathological results. The correlation between relevant radiological signs and MVI status was analyzed. MaZda4.6 software and the mutual information method were employed to identify the top 10 dominant texture features, which were combined with radiological signs to construct artificial neural network (ANN) models for MVI prediction. The predictive performance of the ANN models was evaluated using area under the curve, sensitivity, and specificity. ANN models with relatively high predictive performance were screened using the DeLong test, and the regression model of multilayer feedforward ANN with backpropagation and error backpropagation learning method was used to evaluate the models' stability. The absence of a pseudocapsule, an incomplete pseudocapsule, and the presence of tumor blood vessels were identified as independent predictors of HCC MVI. The ANN model constructed using the dominant features of the combined group (pseudocapsule status + tumor blood vessels + arterial phase + venous phase) demonstrated the best predictive performance for MVI status and was found to be automated, highly operable, and very stable. The ANN model constructed using the dominant features of the combined group can be recommended as a non-invasive method for preoperative prediction of HCC MVI status.

Feature optimized hybrid model for prediction of myocardial infarction

Background Cardiovascular disease is rampant worldwide and has become the leading factor in increasing the global mortality rates. According to the World Heart Federation, death toll due to CVD has increased from 12.1 million in 1990 to around 19 million in 2019. Myocardial Infarction (MI) is a condition where the heart muscle dies due to reduced or inhibited flow of oxygenated blood. It has affected approximately 3 million people worldwide, with more than 1 million deaths in the United States annually. Such unusual proliferation in global death toll due to CVD can be reduced to a great extent by predicting the risk of CVD at an early stage. Method In this paper, several feature selection techniques including Variance-based, Mutual Information (MI), Maximum Relevance Minimum Redundancy (MRMR), Boruta, and Recursive Feature Elimination (RFE) algorithms are used feature optimization. For class prediction, the Logistic Regression (LR), Support Vector Machine (SVM), Decision Tree (DT), and Adaboost algorithms were implemented in their ordinary, One-vs-Rest (OVR) and One-vs-One (OVO) methods. Result The performance of Adaboost model has significantly improved by using feature selection techniques, that is, the accuracy of 74% (without any feature selection taking 5.3 seconds) is increased to 85% (with Boruta feature selection taking only 2.17 seconds training time) and 88% (with MRMR feature selection taking 1.6 seconds training time). Similarly, the DT-OVO model’s performance has improved from 84% (without any feature selection taking 1.48 seconds training time) to 86% (with Boruta feature selection taking 0.58 training time). For other models, the performance is maintained with reduced model training times. Conclusion This research paper prioritizes on feature selection in developing machine learning models for CVD prediction. This conclusion is justified by demonstrating the significant reduction in model training times for the 72 models generated while maintaining or even improving the model’s predictive performance.