Leave-one-out Cross-validation Research Articles

Human movement detection and classification capabilities using passive Wi-Fi based radar

Human detection and classification via Wi-Fi transmission have received a lot of attention in recent years as crucial facilitators in security and human-computer interaction (HCI). The passive Wi-Fi radar (PWR) system used by previous researchers applied cross-ambiguity function (CAF) and CLEAN algorithms to process the detected signals. This paper explores the feasibility and viability of a PWR system in detecting and classifying human movements without utilizing CAF and CLEAN algorithms. The movements are performed by four participants but with comparable body sizes and heights. Three daily human movements are investigated namely walking, bending, and sitting, with each participant performing each movement 24 times, providing a total of 96 samples per activity. The system is evaluated based on the consistency of the signal pattern in a frequency domain and the percentage accuracy is assessed using an artificial neural network (ANN) classifier and trained using a leave-one-out cross-validation (LOOCV) method. The frequency domain results reveal that the signals are consistent, with no noticeable variations or changes in the voltage intensity or shape of the main lobe. The classification of the movements shows that the classifier has an overall accuracy of 97.6%.

Modelo estadístico de la descontaminación de aguas residuales que contienen Carbendazim mediante foto-Fenton

Homogeneous photo-Fenton photocatalysis is one of the most widely used advanced oxidation processes in treating wastewater containing pesticides, where optimizing the catalyst dosage and oxidizing agent, with the response being mineralization in terms of total organic carbon (TOC) or removal of the specific contaminant, is the goal of any experimental design. Experimental development in treatment processes necessitates executing a significant number of experimental conditions that require the use of reagents, energy, and execution time. Therefore, modeling such phenomena emerges as an alternative to these limitations in wastewater treatment. In this research, the influence of factors FeSO4 and H2O2, each at three levels, on the mineralization of a synthetic solution of the insecticide Carbendazim (50 mg/L) in terms of TOC was evaluated using a multiple linear regression model optimized by response surface methodology. The main results established that the best model fit considers the interaction between FeSO4 and H2O2 (X1*X2) and the quadratic terms of each (X21,X22) with p-values < 0.05. The validation of the model using the Leave-One-Out Cross Validation (LOOCV) technique, as well as accuracy and precision through residual analysis and ordinary least squares assumptions, confirms the validity of the conclusions derived from it.

Comparison of machine learning and deep learning models for the assessment of rondo wine grape quality with a hyperspectral camera

Hyperspectral images provide rich spectral/spatial data and have shown remarkable performance in precision viticulture. Standardized data-processing methods are necessary to reduce the dimensionality and to identify powerful wavelengths. Toward this goal, we evaluated and compared the performance and novel-wavelength-identification ability of five renowned machine learning models: the linear models Ridge and LASSO, a 1D + 2D convolutional neural network (1D +2D CNN) non-linear model, a gradient boosting decision tree (GBDT) using XGBoost as an ensemble model, and an explainable boosting machine (EBM) followed by support vector regression (SVR) as a hybrid model. The model evaluations were conducted using leave-one-out cross-validation (LOOCV) as we sought to clarify the best-fitted machine learning model. Our results demonstrated that Ridge, LASSO, showed better performance with relatively high accuracy but were weak as a wavelength identifier. GDBT-XGBoost showed considerable prediction power and wavelength identification. EBM-SVM emerged as the most powerful model, demonstrating exceptional stability and clear wavelength classification even for destructive measurements under varying environmental stresses across Rondo's growth stages. The combined approach of 1D + 2D CNN algorithms was advantageous to handle the dynamic shapes of spectral curve and horizontal shift of the wavelengths obtained from outdoor data acquisition, and notably, it showed the highest accuracy to predict the brix and pH of wine grapes for both indoor and outdoor sensings. But the combined effects of 1D and 2D CNN algorithms were difficult to clarify the importance of spectral features for the brix and pH prediction. The integrated machine learning models with dimensionality reduction, and proper image acquisition can increase the model's accuracy. The common absorption peaks were observed in the near-infrared region of 700 nm and 900 nm. Those wavelengths should be considered for the development of low-cost sensing platforms with fewer bands. Wavelengths over 900 nm are also important to develop outdoor sensing platforms.

Utility of CT Radiomics and Delta Radiomics for Survival Evaluation in Locally Advanced Nasopharyngeal Carcinoma with Concurrent Chemoradiotherapy.

A high incidence rate of nasopharyngeal carcinoma (NPC) has been observed in Southeast Asia compared to other parts of the world. Radiomics is a computational tool to predict outcomes and may be used as a prognostic biomarker for advanced NPC treated with concurrent chemoradiotherapy. Recently, radiomic analysis of the peripheral tumor microenvironment (TME), which is the region surrounding the gross tumor volume (GTV), has shown prognostic usefulness. In this study, not only was gross tumor volume (GTVt) analyzed but also tumor peripheral regions (GTVp) were explored in terms of the TME concept. Both radiomic features and delta radiomic features were analyzed using CT images acquired in a routine radiotherapy process. A total of 50 patients with NPC stages III, IVA, and IVB were enrolled between September 2004 and February 2014. Survival models were built using Cox regression with clinical factors (i.e., gender, age, overall stage, T stage, N stage, and treatment dose) and radiomic features. Radiomic features were extracted from GTVt and GTVp. GTVp was created surrounding GTVt for TME consideration. Furthermore, delta radiomics, which is the longitudinal change in quantitative radiomic features, was utilized for analysis. Finally, C-index values were computed using leave-one-out cross-validation (LOOCV) to evaluate the performances of all prognosis models. Models were built for three different clinical outcomes, including overall survival (OS), local recurrence-free survival (LRFS), and progression-free survival (PFS). The range of the C-index in clinical factor models was (0.622, 0.729). All radiomics models, including delta radiomics models, were in the range of (0.718, 0.872). Among delta radiomics models, GTVt and GTVp were in the range of (0.833, 0.872) and (0.799, 0.834), respectively. Radiomic analysis on the proximal region surrounding the gross tumor volume of advanced NPC patients for survival outcome evaluation was investigated, and preliminary positive results were obtained. Radiomic models and delta radiomic models demonstrated performance that was either superior to or comparable with that of conventional clinical models.

Machine learning based technique to predict the water adulterant in milk using portable near infrared spectroscopy

Milk adulteration is a significant problem globally, as it is the most widely consumed and essential food product. Due to this, monitoring milk quality is necessary for sustaining human health. A Machine Learning (ML) based non-destructive system was developed to identify water adulteration in milk using Near Infrared (NIR) Spectroscopy. A database was created by mixing water in milk in varying proportions (0–40 %) and capturing spectra using compact TI DLP NIR scan Nano spectroscopy in the 900–1700 nm range. The captured spectra were pre-processed with the Savitzky-Golay (SG) filter, Multiplicative Scatter Correction (MSC), and Standard Normal Variate (SNV) method. The most informative wavelength points were selected using the wavelength/feature selection technique, and the dimensions of these wavelengths were reduced using Principal Component Analysis (PCA). Various ML models were employed to predict the water concentration in milk. Both classification and regression methods were applied to check the system's performance. In the regression analysis, the k-Nearest Neighbour (KNN) achieved the best R2, Root Mean Square Error (RMSE), Standard Error of Prediction (SEP), Mean Absolute Error (MAE), Ratio of Performance to Deviation (RPD), Leave One Out Cross-Validation (LOOCV)-R2, and LOOCV-RMSE of 0.999, 0.399 mL ( % v/v), 0.096 mL ( % v/v), 0.227 mL ( % v/v), 33.005, 0.999, and 0.353 mL ( % v/v), respectively, while for classification analysis, the Random Forest (RF) achieved 100 % accuracy and Matthew's Correlation Coefficient (MCC). The impact of the proposed portable system, which has the potential to reshape practices and set new standards for food quality assurance, transformative, and offers solutions to critical challenges in the dairy industry.

Utilizing CNN to predict homogeneous thermo-mechanical properties of conductive layers for reliability numerical analysis in electronics

This study explores the application of Convolutional Neural Networks (CNN) in predicting the partitioned homogeneous properties (PHPs) of electronic product wiring structures, aiming to enhance the efficiency of reliability analysis through Finite Element Analysis (FEA). A systematic and novel method was developed to generate the input partitioned wiring diagram image sets that foster model generalization and universality. The performance of the CNN-based approach was assessed through regression analysis by employing a leave-one-out cross-validation (LOOCV) approach across three PCBs. Additionally, the predicted PHPs of one of the PCBs were applied in product-level FEA to validate their reliability, further demonstrating the practical applicability of the CNN-based method. The results demonstrate that a well-trained CNN model can accurately predict the properties of previously unencountered wiring structures, thereby facilitating direct application in product-level reliability FEA and improving the efficiency of reliability assessment. Furthermore, efficiency evaluation revealed that the CNN-based method offers significant advantages in terms of time and cost economy compared to the mesoscopic FEA method in determining, highlighting its potential for broader application in electronic product reliability analysis. The findings provide preliminary insights and propose strategies to enhance the applicability of CNN methods in this domain, ultimately aiming to improve the efficiency and reduce costs in reliability assessments, thereby streamlining the overall product development process.

Design and validation of a diagnostic suspicion checklist to differentiate epileptic from psychogenic nonepileptic seizures (PNES-DSC)

ObjectivePsychogenic non-epileptic seizures (PNES) are complex clinical manifestations and misdiagnosis as status epilepticus remains high, entailing deleterious consequences for patients. Video-electroencephalography (vEEG) remains the gold-standard method for diagnosing PNES. However, time and economic constraints limit access to vEEG, and clinicians lack fast and reliable screening tools to assist in the differential diagnosis with epileptic seizures (ES). This study aimed to design and validate the PNES-DSC, a clinically based PNES diagnostic suspicion checklist with adequate sensitivity (Se) and specificity (Sp) to discriminate PNES from ES. MethodsA cross-sectional study with 125 patients (n = 104 drug-resistant epilepsy; n = 21 PNES) admitted for a vEEG protocolised study of seizures. A preliminary PNES-DSC (16-item) was designed and used by expert raters blinded to the definitive diagnosis to evaluate the seizure video recordings for each patient. Cohen's kappa coefficient, leave-one-out cross-validation (LOOCV) and balance accuracy (BAC) comprised the main validation analysis. ResultsThe final PNES-DSC is a 6-item checklist that requires only two to be present to confirm the suspicion of PNES. The LOOCV showed 71.4% BAC (Se = 45.2%; Sp = 97.6%) when the expert rater watched one seizure video recording and 83.4% BAC (Se = 69.6%; Sp = 97.2%) when the expert rater watched two seizure video recordings. ConclusionThe PNES-DSC is a straightforward checklist with adequate psychometric properties. With an integrative approach and appropriate patient history, the PNES-DSC can assist clinicians in expediting the final diagnosis of PNES when vEEG is limited. The PNES-DSC can also be used in the absence of patients, allowing clinicians to assess seizure recordings from smartphones.

Design and validation of novel maize grain cleaning loss detection system based on classification models of particle time-domain signals

Quantitative evaluation of grain cleaning loss is a central concern in maize intelligent direct harvesting. This study integrates particle time-domain signals with machine learning models to develop a novel grain cleaning loss detection system and create modular embedded companion software. Specifically, collisional time-domain signals of grains and impurities with varying mass gradients are captured, and abnormal data are eliminated utilizing the Mahalanobis Distance (MD) algorithm. Signal classification models incorporating diverse preprocessing methods are trained using Naive Bayes (NB) and Support Vector Machine (SVM) algorithms. Model performance is evaluated through internal leave-one-out-cross-validation (LOOCV) and external validation sets. The best classification model is integrated into the grain loss detection software, and bench tests verify real-time cleaning loss rates for detection systems. Results demonstrate that the SVM model with Savitzky-Golay (S-G) smoothing + Second-Order Derivative (SOD) preprocessing achieves the highest predictive accuracy, with 97.44 % and 96.15 % in training and validation sets, respectively. When the feeding rate of maize mixtures is 5 kg/s, absolute grain loss rates for three measurements are 1.62 %, 1.72 % and 1.86 %, respectively. Compared to traditional loss detection systems with multifunctional circuits, the novel loss detection system enhances the average accuracy of grain loss rate by 6.54, 6.82 and 8.93 percentage points. Hence, the proposed method and apparatus can be applied to detect grain cleaning loss quantitatively.

Hologram quantitative structure-activity relationship on the gas chromatographic retention index of plant essential oil constituents

The gas chromatography retention index (RI) is an important parameter for the identification of different types of compounds in the field of chromatographic analysis; however, the experimental collection of RI values is a extremely cumbersome process. Thus, there is an urgent need for the establishment of a simple, efficient, and accurate model for the prediction of the RI values of compounds. In this study, first, the experimental RI values for 60 plant essential oil constituents were obtained. Next, a model describing the hologram quantitative structure-activity relationship (HQSAR) between the structural properties of the essential oil constituents and their RI values was investigated and constructed. The optimal HQSAR model was established by setting the model parameters "fragment size", "fragment distinction", "hologram length" and "principal components" to "1-4", "C, Ch", "199", and "4", respectively. Finally, the predictive ability of the model was verified using external test set validation and leave-one-out cross-validation (LOO-CV). The experimental results were as follows, the root mean square error of prediction (RMSEP), predictive determination coefficient ([Formula: see text]), concordance correlation coefficient (CCC), and mean relative error (MRE) for external test set validation were 40.45, 0.984, 0.968, and 2.20%, respectively. Meanwhile, the root mean square error of cross validation (RMSECV) and MRE for LOO-CV were 72.56 and 4.17%, respectively. These findings demonstrate that the established HQSAR model has a good predictive ability and can accurately predict the RI values of plant essential oil constituents. In addition, the molecular contribution maps of the HQSAR model revealed that the RI values of aromatic compounds increase when hydroxyl groups are connected to their alkyl chains. Aliphatic compounds feature long chain alkyl groups, which can lead to an increase in RI values. The above phenomena highlight the promising application prospects of HQSAR for studying the RI values of plant essential oil constituents. Therefore, this study provides a reliable theoretical basis for predicting the RI values of other essential oil constituents.

Comprehensive analysis of the interaction of antigen presentation during anti-tumour immunity and establishment of AIDPS systems in ovarian cancer.

There are hundreds of prognostic models for ovarian cancer. These genes are based on different gene classes, and there are many ways to construct the models. Therefore, this paper aims to build the most stable prognostic evaluation system known to date through 101 machine learning strategies. We combined 101 algorithm combinations with 10 machine learning algorithms to create antigen presentation-associated genetic markers (AIDPS) with outstanding precision and steady performance. The inclusive set of algorithms comprises the elastic network (Enet), Ridge, stepwise Cox, Lasso, generalized enhanced regression model (GBM), random survival forest (RSF), supervised principal component (SuperPC), Cox partial least squares regression (plsRcox), survival support vector machine (Survival-SVM). Then, in the train cohort, the prediction model was fitted using a leave-one cross-validation (LOOCV) technique, which involved 101 different possible combinations of prognostic genes. Seven validation data sets (GSE26193, GSE26712, GSE30161, GSE63885, GSE9891, GSE140082 and ICGC_OV_AU) were compared and analysed, and the C-index was calculated. Finally, we collected 32 published ovarian cancer prognostic models (including mRNA and lncRNA). All data sets and prognostic models were subjected to a univariate Cox regression analysis, and the C-index was calculated to demonstrate that the antigen presentation process should be the core criterion for evaluating ovarian cancer prognosis. In a univariate Cox regression analysis, 22 prognostic genes were identified based on the expression profiles of 283 genes involved in antigen presentation and the intersection of genes (p < 0.05). AIDPS were developed by our machine learning-based integration method, which was applied to these 22 genes. One hundred and one prediction models are fitted using the LOOCV framework, and the C-index is calculated for each model across all validation sets. Interestingly, RSF + Lasso was the best model overall since it had the greatest average C-index and the highest C-index of any combination of models tested on the validated data sets. In comparing external cohorts, we found that the C-index correlated AIDPS method using the RSF + Lasso method in 101 prediction models was in contrast to other features. Notably, AIDPS outperformed the vast majority of models across all data sets. Antigen-presenting anti-tumour immune pathways can be used as a representative gene set of ovarian cancer to track the prognosis of patients with cancer. The antigen-presenting model obtained by the RSF + Lasso method has the best C-INDEX, which plays a key role in developing antigen-presenting targeted drugs in ovarian cancer and improving the treatment outcome of patients.

Abstract 2986: Germline signatures predicting toxicity and response to CTLA4 inhibitors

Abstract Background: Because germline microRNA-based variants predict toxicity risk to anti-PD-1/PDL-1 treatment, and CTLA4 inhibitors are often used in combination with them, we wanted to see if we could also predict toxicity and/or response to anti-CTLA4 therapy. Methods: The study included 69 patients from 3 institutions treated with anti-CTLA4 therapy alone. Patients were 23-89 years old with median age of 62. “Toxicity” was defined as grade 3 or higher toxicity and “Response” as a complete response, partial response, or stable disease. A panel of previously tested variants (n=139), age, sex and cycles of anti-CTLA4 were used to predict endpoints of interest. To balance model dimensionality we pre-screened variants and included them if they were marginally associated with our outcomes of interest via Fisher’s Exact Test or Jonckheere-Terpstra Test at a p-value threshold of &amp;lt; 0.2. We fit Elastic Net (EN), Random Forest (RF), and Boosted Tree (BT) models with up-sampling to predict outcomes of interest. Leave-one-out cross validation (LOOCV) was used to evaluate model performance, and the model with the highest LOOCV AUC (between all EN, RF, and BT) was chosen as the final model and re-fit using all observations. We also performed a stratified gene ontology (GO) analysis to assess biological differences between variants involved in the signatures. Our GO analysis compared gene clusters across groupings with an adjusted p-value cutoff of 0.05 against a universal genomic background. Results: 15 patients had toxicity and 21 patients had a response, with 27/139 and 25/139 variants marginally associated with these endpoints, respectively. Our toxicity model had a LOOCV AUC of 0.811 with the top variables being age, TNNT2_rs3729843, REV3L_rs465646, miR34b_c_rs4938723, and XRCC3_rs861539. Our response model had an LOOCV AUC of 0.716 with top variants being IL2RA_rs2476491, REV3L_rs465646, and IL10RB_rs2834167. Including anti-CTLA4 cycles improved the prediction accuracy for Response with an AUC 0.764. GO analysis revealed 117 pathways enriched in both the Toxicity and Response signatures, which were predominantly involved in pri-miRNA transcriptional regulation. Top GO terms in the Response signature (333 unique) were related to apoptosis and cell death. Top GO terms in the Toxicity signature (425 unique) were related to DNA replication and cellular signaling and response, including the interleukin-35-mediated signaling pathway. Conclusion: We identified germline signatures predicting toxicity and response following anti-CTLA4 therapy. These signatures differed from each other as well as from a previously identified signature of toxicity to anti-PD1/PDL-1 therapy, and GO analysis identified plausible underlying signaling pathways. Our findings continue to validate the potential of these variants to predict meaningful endpoints in cancer therapy, and thus further research in larger and more diverse cohorts is warranted. Citation Format: Joanne B. Weidhaas, Kristen McGreevy, Nicholas Marco, Nora Sundahl, Christopher R. Cabanski, Christine Spencer, Piet Ost, Donatello Telesca. Germline signatures predicting toxicity and response to CTLA4 inhibitors [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 2986.

Honey differentiation with FTIR-ATR spectroscopy – Comparison with physicochemical parameters of a Polish honey sample set

The varietal discrimination of honey is important because its bioactive properties, as well as specific characteristics, depend on the botanical origin. The present study aimed to verify the discriminatory power of Fourier-transform infrared spectroscopy (FT-IR) for differentiating Polish honeys and compare the results with the discrimination power of a model built on the basis of selected physicochemical parameters. For this purpose, buckwheat, rape, acacia, heather, goldenrod, and coniferous honeydew honeys were collected from trustworthy apiaries in different regions of Poland. The samples were further authenticated using pollen analysis and tested by means of FT-IR and the attenuated total reflection (ATR) technique. Basic psychochemical parameters were also assessed. The obtained data were analyzed by means of principal component analysis (PCA) and linear discriminant analysis (LDA). The obtained models were validated with a leave-one-out cross-validation (LOOCV) procedure. PCA analysis confirmed the usefulness of the ATR FT-IR method as a honey differentiation tool. The LDA for the FT-IR model allowed to correctly classify 26 out of 32 samples (81.3% accuracy), which was only slightly lower than for data the model based on the physiochemical characteristics (27 out of 32 samples and 84.4% accuracy), revealing a certain potential of FT-IR as fast evaluation method.

Abstract 4920: A nanopore sequencing approach characterizes cell-free DNA methylation-fragmentomics profiles indicative of breast cancer in a large cohort

Abstract The detection of breast cancer from cell-free DNA (cfDNA) offers a new approach with the potential to benefit individual screening and early detection. However, identifying breast cancer from cfDNA remains a major challenge when one uses conventional short-read sequencing. Methylation and DNA fragmentation characteristics provide highly informative biomarkers for detecting and monitoring cancer. To characterize these genomic features from cfDNA, we developed an approach using single molecule Nanopore sequencing. Single cfDNA molecules are sequenced intact with the methylation and fragment size directly extrapolated from Nanopore data. Moreover, our approach requires small amounts of cfDNA to discern specific genomic features from individual molecules. Thus, the use of Nanopore sequencing provides both methylation and fragment size as directly measured from native cfDNA without the use of any chemical processing or other molecular manipulation. We studied an extended breast cancer cohort with in total 1080 of samples, consisting of 440 untreated breast cancer patients and 640 healthy controls. Nanopore sequencing analysis of these cfDNA samples identified a range of 1 to 4 million CpG sites from 1 ml of plasma. Leveraging the cohort, we developed an ensemble classifier for breast cancer detection based on cfDNA methylation and fragmentation. The classifier leveraged a leave-one-out cross-validation (LOOCV) to generate predicted probability. We created nucleosome-specific methylation scores as the input features to the classifier, which scores quantified the similarity of each sample's methylation profile to a training set on a selected set of CpG sites associated with cancers. Furthermore, we also identified DNA fragmentation features, revealing distinct mono-nucleosome and di-nucleosome patterns that distinguish cancers from healthy controls. Aggregating the DNA nucleosome-specific scores and fragmentation features, we built an ensemble LOOCV random forest model. We applied the model to the breast cancer cohort and achieved a detection specificity of 74%, a sensitivity of 71%, and an overall accuracy of approximately 73%. The area under the receiver operating characteristic curve (AUROC) reached 80%. We validated the classifier with two external independent cohorts. Validation cohort 1 demonstrated as a specificity of 85%, a sensitivity of 80%, and overall accuracy of 83% with AUROC of 90%. Validation cohort 2 without controls demonstrated that the classifier achieved a sensitivity of 94%. Both validation sets confirm that our model can achieve excellent performance for early breast cancer detection. In summary, our findings demonstrate that nanopore sequencing provides cfDNA methylation and fragment profiling that enables breast cancer detection. This approach holds the potential to improve cancer diagnosis and treatment outcomes. Citation Format: Xiangqi Bai, Billy T. Lau, Alison Almada, Sue Grimes, Tianqi Chen, Hojoon Lee, Hanlee P. Ji. A nanopore sequencing approach characterizes cell-free DNA methylation-fragmentomics profiles indicative of breast cancer in a large cohort [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 4920.

Optimizing echo state networks for continuous gesture recognition in mobile devices: A comparative study

Continuous gesture recognition can be used to enhance human-computer interaction. This can be accomplished by capturing human movement with the use of the Inertial Measurement Units in smartphones and using machine learning algorithms to predict the intended gestures. Echo State Networks (ESNs) consist of a fixed internal reservoir that is able to generate rich and diverse nonlinear dynamics in response to input signals that capture temporal dependencies within the signal. This makes ESNs well-suited for time series prediction tasks, such as continuous gesture recognition. However, their application has not been rigorously explored, with regard to gesture recognition. In this study, we sought to enhance the efficacy of ESN models in continuous gesture recognition by exploring diverse model structures, fine-tuning hyperparameters, and experimenting with various training approaches. We used three different training schemes that used the Leave-one-out Cross-validation (LOOCV) protocol to investigate the performance in real-world scenarios with different levels of data availability: Leaving out data from one user to use for testing (F1-score: 0.89), leaving out a fraction of data from all users to use in testing (F1-score: 0.96), and training and testing using LOOCV on a single user (F1-score: 0.99). The obtained results outperformed the Long Short-Term Memory (LSTM) performance from past research (F1-score: 0.87) while maintaining a low training time of approximately 13 seconds compared to 63 seconds for the LSTM model. Additionally, we further explored the performance of the ESN models through behaviour space analysis using memory capacity, Kernel Rank, and Generalization Rank. Our results demonstrate that ESNs can be optimized to achieve high performance on gesture recognition in mobile devices on multiple levels of data availability. These findings highlight the practical ability of ESNs to enhance human-computer interaction.

Qingxiangxing Baijiu sensory quality grade classification by 1H NMR and GC combined with multivariate statistical analysis

The traditional uniform artificial sensory evaluation makes it difficult to standardize the classification of different sensory quality grades of Baijiu. In this study, a total of 92 authentic Qingxiangxing Baijiu samples with 3 sensory quality grades were carefully collected. Gas chromatography (GC) was used to determine 46 main flavor components and proton nuclear magnetic resonance (1H NMR) spectroscopy was employed to obtain hydrogen atom characteristic information of organic compounds. The principal component analysis (PCA), k-nearest neighbor (KNN) and linear discriminant analysis (LDA) models were conducted and fully validated by internal leave-one-out cross validation (LOOCV) and external repeated double random cross validation (RDRCV). The sensory quality grades of Qingxiangxing Baijiu were effectively classified by using GC and 1H NMR techniques coupled with PCA/KNN analysis with the averaged accuracy higher than 80%. In addition, synthetic minority oversampling technique (SMOTE) algorithm was successfully used to address the model overfitting problem caused by an unbalanced sample composition. This study demonstrated that 1H NMR and GC combined with multivariate statistical analysis were effective for sensory quality classification of Qingxiangxing Baijiu.

Automated volumetry of meningiomas in contrast-enhanced T1-Weighted MRI using deep learning

Background: Meningiomas are among the most common intracranial tumors. In these tumors, volumetric assessment is not only important for planning therapeutic intervention but also for follow-up examination.However, a highly accurate automated volumetric method for meningiomas using single-modality magnetic resonance imaging (MRI) has not yet been reported. Here, we aimed to develop a deep learning-based automated volumetry method for meningiomas in MRI and investigate its accuracy and potential clinical applications. Methods: For deep learning, we used MRI images of patients with meningioma who were referred to Osaka University Hospital between January 2007 and October 2020. Imaging data of eligible patients were divided into three non-overlapping groups: training, validation, and testing. The model was trained and tested using the leave-oneout cross-validation method. Dice index (DI) and root mean squared percentage error (RMSPE) were measured to evaluate the model accuracy. Result: A total of 178 patients (64.6 ± 12.3 years [standard deviation]; 147 women) were evaluated. Comparison of the deep learning model and manual segmentation revealed a mean DI of 0.923 ± 0.051 for tumor lesions. For total tumor volume, RMSPE was 9.5 ± 1.2%, and Mann–Whitney U test did not show a significant difference between manual and algorithm-based measurement of the tumor volume (p = 0.96). Conclusion: The automatic tumor volumetry algorithm developed in this study provides a potential volume-based imaging biomarker for tumor evaluation in the field of neuroradiological imaging, which will contribute to the optimization and personalization of treatment for central nervous system tumors in the near future

CircRNA-disease association prediction with an improved unbalanced Bi-Random walk

Studies show that circular RNAs (circRNAs), a type of non-coding RNAs, play various roles in biological processes such as the formation and progression of many different diseases. For this reason, identifying potential circRNAs associated with diseases is vital for early diagnosis. Determining these relationships experimentally requires a long process and is also expensive. For this reason, computational models are being developed to determine the relationships between circRNA and diseases. In this study, we recommend a technique called Improved Unbalanced Bi-Random Walk (UBRW) to identify potential circRNAs associated with diseases. The commonly used 5-fold cross-validation (CV) technique and leave one-out cross-validation (LOOCV) technique were applied to verify the predictive ability of our technique. The area under curve (AUC) values calculated in 5-fold CV and LOOCV are 0.8910 and 0.9669, respectively. Case studies on the commonly occurring gastric cancer and breast cancer were conducted to further validate the predictive performance of our method. When the results were examined, it was seen that the prediction ability of the UBRW method was quite successful.

Diagnosis and classification of intestinal diseases with urine by surface-enhanced Raman spectroscopy

Intestinal Disease (ID) is often characterized by clinical symptoms such as malabsorption, intestinal dysfunction, and injury. If treatment is not timely, it will increase the risk of cancer. Early diagnosis of ID is the key to cure it. There are certain limitations of the conventional diagnostic methods, such as low sensitivity and specificity. Therefore, development of a highly sensitive, non-invasive diagnostic method for ID is extremely important. Urine samples are easier to collect and more sensitive to changes in biomolecules than other pathological diagnostic samples such as tissue and blood. In this paper, a diagnostic method of ID with urine by surface-enhanced Raman spectroscopy (SERS) is proposed. A classification model between ID patients and healthy controls (HC) and a classification model between different pathological types of ID (i.e., benign intestinal disease (BID) and colorectal cancer (CRC)) are established. Here, 830 urine samples, including 100 HC, 443 BID, and 287 CRC, were investigated by SERS. The ID/HC classification model was developed by analyzing the SERS spectra of 150 ID and 100 HC, while BID/CRC classification model was built with 300 BID and 150 CRC patients by principal component analysis (PCA)-support vector machines (SVM). The two established models were internally verified by leave-one-out-cross-validation (LOOCV). Finally, the BID/CRC classification model was further evaluated by 143 BID and 137 CRC patients as an external test set. It shows that the accuracy of the classification model validated by the LOOCV for ID/HC and BID/CRC is 86.4% and 85.56%, respectively. And the accuracy of the BID/CRC classification model with external test set is 82.14%. It shows that high accuracy can be achieved with these two established classification models. It indicates that ID patients in the general population can be identified and BID and CRC patients can be further classified with measuring urine by SERS. It shows that the proposed diagnostic method and established classification models provide valuable information for clinicians to early diagnose ID patients and analyze different stages of ID.

Raman spectroscopy combined with partial least squares (PLS) based on hybrid spectral preprocessing and backward interval PLS (biPLS) for quantitative analysis of four PAHs in oil sludge

Polycyclic aromatic hydrocarbons (PAHs) contained in a large amount of oily sludge produced in petroleum and petrochemical production has become one of the main environmental protection concerns in the industry. The accurate determination of PAHs is of great significance in the field of petroleum geochemistry and environmental protection. In this study, Raman spectroscopy combined with partial least squares (PLS) based on different hybrid spectral preprocessing methods and variable selection strategies was proposed for quantitative analysis of phenanthrene, fluoranthrene, fluorene and naphthalene (Phe, Flt, Flu and Nap) in oil sludge. At first, PAHs in oily sludge was extracted by solid–liquid extraction with methanol as extractant, and Raman spectra of 21 oily sludge samples were collected by portable Raman spectrometer. And then, the influence of first derivative (D1st), wavelet transform (WT) and their hybrid spectral preprocessing on the predictive performance of the PLS calibration model was discussed. Thirdly, biPLS (backward interval partial least squares) was used to optimize the input variables before and after the hybrid spectral preprocessing methods, and the influence of biPLS and the hybrid spectral preprocessing sequence on the predictive performance of the PLS calibration model was discussed. Finally, the predictive performance of the PLS calibration model was optimized according to the results of leave-one-out cross-validation (LOOCV) method. The results show that the biPLS-D1st-WT-PLS calibration model established by using biPLS first to select the characteristic variables, followed by hybrid spectral preprocessing of the characteristic variables, has better prediction performance for Flt (determination coefficient of prediction (R2P) = 0.9987, and the mean relative error of prediction (MREP) = 0.0606). For Phe, Flu and Nap, the WT-biPLS-PLS calibration model has a better predictive effect (R2P are 0.9995, 0.9996 and 0.9983, and MREP are 0.0426, 0.0719 and 0.0497, respectively). In general, portable Raman spectroscopy combined with PLS calibration model based on different hybrid spectral preprocessing and variable selection strategies has achieved good prediction results for quantitative analysis of four PAHs in oily sludge. It is a new strategy to firstly select the characteristic variables of the original spectra, and secondly to preprocess the characteristic variables by the hybrid spectral preprocessing, which will provide a new idea for the establishment of quantitative analysis methods for PAHs in oily sludge.

Harsh, unpredictable childhood environments are associated with inferior frontal gyrus connectivity and binge eating tendencies in late adolescents

Harsh, unpredictable childhood environments (HUCE) are associated with obesity older in life, but knowledge of how HUCE affect binge eating tendencies is lacking. Five hundred and one late adolescents aged 16–22 were recruited to finish resting state functional magnetic resonance imaging scan, behavioral measures including retrospective recall of childhood environmental harshness and unpredictability, binge eating tendencies and demographics. Three hundred and seventy-six of participants further completed the computerized visual probe task designed to evaluate attentional engagement towards high and low calorie food. As right inferior frontal gyrus (IFG) was the key nodes that related to both early life adversity and binge eating tendencies, it was treated as the interest region in the dynamic functional connectivity analyses. Results found that HUCE are associated with significant but modest decreases in connectivity of right inferior frontal gyrus (IFG)– bilateral medial frontal gyrus, right IFG – bilateral inferior parietal lobule (IPL), and right IFG – left superior frontal gyrus connectivity, as well as attentional engagement to high-calorie food and binge eating tendencies. A machine-learning method named linear support vector regression (SVR) and leave one out cross-validation (LOOCV) procedure used to examine the robustness of the brain-behavior relationship further confirm the findings. Mediation analyses suggested that right IFG – left IPL connectivity mediates the association of HUCE and binge eating tendencies. Findings suggest right IFG – left IPL connectivity may serve as a crucial neurobiological underpinning of HUCE to regulate binge eating behaviors. As such, these results contribute to a novel perspective and hypotheses in elucidating developmental neuro-mechanisms related to binge eating.