Non-normalized Data Research Articles

Development of fuzzified neural network for enterprise bankruptcy risk estimation

The object of this study is the assessment of the level of enterprise bankruptcy risk. It is a critical component in assessing the financial condition of an enterprise, and also serves as an indicator that allows the management team to reduce potential risks and develop their own strategies to strengthen the financial condition of the enterprise. One of the most challenging aspects of bankruptcy forecasting is the complex financial situations of bankrupt companies. By accurately predicting the risk of bankruptcy, businesses can take preventive measures to mitigate financial difficulties and ensure long-term sustainability. Previous methods, such as Altman's Z-score, are not accurate enough, as presented in the study. The paper investigates a modern approach to bankruptcy prediction based on a neural network with complex neural elements, namely neural arithmetic logic units (NALUs) and a custom phasing layer. This layer can process complex raw numerical values, such as financial indicators relevant to the analysis of a company's bankruptcy. Compared to Altman's Z-score, the developed method demonstrates a better F1 score in bankruptcy classification (48 %). On the raw data, the neural network demonstrates an improvement in the F1 score by about 40 % compared to the classical multilayer perceptron (MLP) with linear layers and nonlinear activation functions. A modern replacement for ReLU called Mish was used, which achieves better generalization. It was also assumed that the addition of new neural elements, which provide the neural network with arithmetic capabilities, contributes to the performance of processing non-normalized input data. This work highlights the importance of using advanced neural network architectures to improve the accuracy and reliability of forecasting in financial risk assessment. Using the parameters presented in the study, managers of enterprises will be able to more accurately assess the risk of bankruptcy.

Corpus-based dialectometry with topic models

Abstract This paper presents a topic modeling approach to corpus-based dialectometry. Topic models are most often used in text mining to find latent structure in a collection of documents. They are based on the idea that frequently co-occurring words present the same underlying topic. In this study, topic models are used on interview transcriptions containing dialectal speech directly, without any annotations or preselected features. The transcriptions are modeled on complete words, on character n-grams, and after automatical segmentation. Data from three languages, Finnish, Norwegian, and Swiss German, are scrutinized. The proposed method is capable of discovering clear dialectal differences in all three datasets, while reflecting the differences between them. The method provides a significant simplification of the dialectometric workflow, simultaneously saving time and increasing objectivity. Using the method on non-normalized data could also benefit text mining, which is the traditional field of topic modeling.

Abstract LB012: Evaluating computational approaches for CPTAC pan-cancer cross-cohort protein expression comparison

Abstract Recently, the National Cancer Institute’s Clinical Proteomic Tumor Analysis Consortium (CPTAC) has generated harmonized genomic, transcriptomic, proteomic, and clinical data for >1000 tumors in 10 cohorts to facilitate pan-cancer discovery research. However, protein expression comparison across CPTAC cohorts remains challenging due to non-uniform missing data and protein expression distribution patterns across tumor types. Here we present our efforts to evaluate various missing data handling and normalization strategies to generate a normalized pan-cancer protein expression dataset. First, we developed a novel algorithm to select robustly expressed proteins in tumors in any of the CPTAC cohorts, Second, we applied a cohort hybrid imputation approach to protein abundance values from FragPipe within each cohort based on protein expression distribution patterns. Third, we calculated iBAQ using protein abundance value and applied global quantile normalization or smooth quantile normalization methods. To assess if our missing data imputation and normalization strategy affected downstream analyses, we compared the fold change in differential protein expression between tumor and matched normal for each cohort using non-normalized, global quantile normalized and smooth quantile normalized protein iBAQ values. Our results demonstrate a strong correlation in fold change between global quantile normalized data and non-normalized data (Pearson r = 0.97 (ccRCC), r = 0.96 (COAD), r = 0.99 (LUAD) and r = 0.99 (LSCC)). Similar results were observed comparing smooth quantile normalized data to non-normalized data (Pearson r = 1.00 for ccRCC, COAD, LUAD, and LSCC), indicating both normalization methods retained biological differences between tumor and matched normal tissues within cohorts. Lastly, we identified several proteins (ERAP2, CA9, GSTM3, MX1, STAT1) whose protein and RNA expression were highly correlated across eight CPTAC cohorts (r > 0.7 for COAD, BRCA, LUAD, ccRCC, PDAC, UCEC, HNSCC, and LSCC). We then compared their protein expression rank across CPTAC cohorts with their RNA expression rank across corresponding TCGA cohorts. Specifically, median log2(iBAQ) of CPTAC and median log2(TPM) of TCGA are calculated for those proteins within indications, then indications are ranked by median log2(iBAQ) and median log2(TPM) in CPTAC and TCGA, respectively. Weighted rank correlation was used to measure rank agreement. Global quantile normalization has the highest rank correlation (weighted rank correlation between 0.597 to 0.931) compared to smooth quantile normalization or without normalization. These results suggest that combination of cohort hybrid imputation and global quantile normalization is a reasonable approach to generate a normalized CPTAC pan-cancer protein dataset that could be leveraged to interrogate protein expression across different cancer types. Citation Format: Jixin Wang, Xiaowen Tian, Wen Yu, John Bullen, Elaine Hurt, Wenyan Zhong. Evaluating computational approaches for CPTAC pan-cancer cross-cohort protein expression comparison [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(7_Suppl):Abstract nr LB012.

Anomaly Detection in Time Series Data Using Reversible Instance Normalized Anomaly Transformer.

Anomalies are infrequent in nature, but detecting these anomalies could be crucial for the proper functioning of any system. The rarity of anomalies could be a challenge for their detection as detection models are required to depend on the relations of the datapoints with their adjacent datapoints. In this work, we use the rarity of anomalies to detect them. For this, we introduce the reversible instance normalized anomaly transformer (RINAT). Rooted in the foundational principles of the anomaly transformer, RINAT incorporates both prior and series associations for each time point. The prior association uses a learnable Gaussian kernel to ensure a thorough understanding of the adjacent concentration inductive bias. In contrast, the series association method uses self-attention techniques to specifically focus on the original raw data. Furthermore, because anomalies are rare in nature, we utilize normalized data to identify series associations and employ non-normalized data to uncover prior associations. This approach enhances the modelled series associations and, consequently, improves the association discrepancies.

Comparison of the Ocular Surface Disease Index and the Symptom Assessment in Dry Eye Questionnaires for Dry Eye Symptom Assessment.

Patient-reported dry eye symptoms (DESs), assessed using the Ocular Surface Disease Index (OSDI) and the Symptom Assessment iN Dry Eye (SANDE) questionnaires, were compared in a large sample of patients. The correlation (Spearman coefficient) and agreement (Bland-Altman analysis) between the OSDI and SANDE questionnaire scores (with and without score normalization) were assessed in 1033 patients and classified according to the OSDI score as non-DES and DES in a cross-sectional analysis. The normalized and non-normalized SANDE results were higher than the OSDI results in all samples (2.83 ± 12.40 (p = 0.063) and 2.85 ± 15.95 (p = 0.016), respectively) and in non-DES (p > 0.063) and DES (p < 0.001) with both OSDI cutoff values. Weak correlations were found (Spearman coefficient <0.53; p < 0.001) in all cases except DES (0.12, p = 0.126). Weak agreement was found with a Bland-Altman analysis of the normalized and non-normalized scores of both questionnaires (mean difference from -7.67 ± 29.17 (DES patients) to -1.33 ± 8.99 (non-DES patients) without score normalization, and from -9.21 ± 26.37 (DES patients) to -0.85 ± 4.01 (non-DES) with data normalization), with a statistically significant linear relationship (R2 > 0.32, p < 0.001). The SANDE questionnaire did not yield the same patient classification as OSDI. The same operative curves (ROC) of the SANDE normalized and non-normalized scores were used to differentiate among patients with DES using OSDI < 12 (0.836 ± 0.015) or OSDI < 22 (0.880 ± 0.015) cutoff values. Normalized and non-normalized data collected from the SANDE questionnaire showed relevant differences from those of the OSDI, which suggests that the results of the SANDE visual analog scale-based questionnaire provide different patient classifications than the OSDI score.

Targeting a free viral fraction enhances the early alert potential of wastewater surveillance for SARS-CoV-2: a methods comparison spanning the transition between delta and omicron variants in a large urban center.

Wastewater surveillance has proven to be a valuable approach to monitoring the spread of SARS-CoV-2, the virus that causes Coronavirus disease 2019 (COVID-19). Recognizing the benefits of wastewater surveillance as a tool to support public health in tracking SARS-CoV-2 and other respiratory pathogens, numerous wastewater virus sampling and concentration methods have been tested for appropriate applications as well as their significance for actionability by public health practices. Here, we present a 34-week long wastewater surveillance study that covers nearly 4 million residents of the Detroit (MI, United States) metropolitan area. Three primary concentration methods were compared with respect to recovery of SARS-CoV-2 from wastewater: Virus Adsorption-Elution (VIRADEL), polyethylene glycol precipitation (PEG), and polysulfone (PES) filtration. Wastewater viral concentrations were normalized using various parameters (flow rate, population, total suspended solids) to account for variations in flow. Three analytical approaches were implemented to compare wastewater viral concentrations across the three primary concentration methods to COVID-19 clinical data for both normalized and non-normalized data: Pearson and Spearman correlations, Dynamic Time Warping (DTW), and Time Lagged Cross Correlation (TLCC) and peak synchrony. It was found that VIRADEL, which captures free and suspended virus from supernatant wastewater, was a leading indicator of COVID-19 cases within the region, whereas PEG and PES filtration, which target particle-associated virus, each lagged behind the early alert potential of VIRADEL. PEG and PES methods may potentially capture previously shed and accumulated SARS-CoV-2 resuspended from sediments in the interceptors. These results indicate that the VIRADEL method can be used to enhance the early-warning potential of wastewater surveillance applications although drawbacks include the need to process large volumes of wastewater to concentrate sufficiently free and suspended virus for detection. While lagging the VIRADEL method for early-alert potential, both PEG and PES filtration can be used for routine COVID-19 wastewater monitoring since they allow a large number of samples to be processed concurrently while being more cost-effective and with rapid turn-around yielding results same day as collection.

DNN model development of biogas production from an anaerobic wastewater treatment plant using Bayesian hyperparameter optimization

Deep neural networks have been regarded as accurate models to predict complex fermentation processes due to their capacity to learn from a high number of data sets via artificial intelligence. To enhance the performance of these models the main challenge is to select the appropriate hyperparameters, such as neural network architecture (number of hidden layers and hidden units), activation function, optimizer, learning rate, and others.This study aims to use a hybrid Bayesian optimization with tree-structured Parzen estimator (BO-TPE) to predict biogas production from real wastewater treatment plant data using deep neural network machine learning (DNN) with optimized hyperparameters. Due to the high number of missing sample measurement records, the data preprocessing process has been performed in three sequential steps: the first step is to remove columns with a high percentage of missing values. The second step concerns removing rows with a high number of missing values. The remaining missing values in the dataset were removed in the third step using the dropna function from Pandas' library. Then, on the remaining data, three normalization techniques (MinMaxScaler, RobustScaler, and StandardScaler) were used for 16 relevant features of the anaerobic digestion process (AD) and compared with Non-Normalized data. The RobustScaler technique demonstrated good prediction capabilities of biogas volume produced. The maximum predicted volume was 2236.105 Nm3/d, with the coefficient of determination (R2), the mean absolute error (MAE), and the root mean square error (RMSE) values of 0.712, 164.610, and 223.429, respectively. This value remains slightly higher than the actual biogas volume produced in the wastewater treatment plant, which was 2131 Nm3/d. Adopting StandardScaler, MinMaxScaler, and Non-Normalized data provided statistical performance indicator values of (0.690; 170.518; 231.519), (0.679; 180.575; 235.648), and (0.647; 183.403; 247.203), respectively. These findings offer an autonomous strategy to monitor the effective operational variables of a large-scale digester, ensuring the stability of the complex fermentation process as well as the long-term sustainability and economic viability of the renewable energy approach in the sector of waste treatment.

The use of nonnormalized surface EMG and feature inputs for LSTM-based powered ankle prosthesis control algorithm development.

Advancements in instrumentation support improved powered ankle prostheses hardware development. However, control algorithms have limitations regarding number and type of sensors utilized and achieving autonomous adaptation, which is key to a natural ambulation. Surface electromyogram (sEMG) sensors are promising. With a minimized number of sEMG inputs an economic control algorithm can be developed, whereas limiting the use of lower leg muscles will provide a practical algorithm for both ankle disarticulation and transtibial amputation. To determine appropriate sensor combinations, a systematic assessment of the predictive success of variations of multiple sEMG inputs in estimating ankle position and moment has to conducted. More importantly, tackling the use of nonnormalized sEMG data in such algorithm development to overcome processing complexities in real-time is essential, but lacking. We used healthy population level walking data to (1) develop sagittal ankle position and moment predicting algorithms using nonnormalized sEMG, and (2) rank all muscle combinations based on success to determine economic and practical algorithms. Eight lower extremity muscles were studied as sEMG inputs to a long-short-term memory (LSTM) neural network architecture: tibialis anterior (TA), soleus (SO), medial gastrocnemius (MG), peroneus longus (PL), rectus femoris (RF), vastus medialis (VM), biceps femoris (BF) and gluteus maximus (GMax). Five features extracted from nonnormalized sEMG amplitudes were used: integrated EMG (IEMG), mean absolute value (MAV), Willison amplitude (WAMP), root mean square (RMS) and waveform length (WL). Muscle and feature combination variations were ranked using Pearson's correlation coefficient (r > 0.90 indicates successful correlations), the root-mean-square error and one-dimensional statistical parametric mapping between the original data and LSTM response. The results showed that IEMG+WL yields the best feature combination performance. The best performing variation was MG + RF + VM (rposition = 0.9099 and rmoment = 0.9707) whereas, PL (rposition = 0.9001, rmoment = 0.9703) and GMax+VM (rposition = 0.9010, rmoment = 0.9718) were distinguished as the economic and practical variations, respectively. The study established for the first time the use of nonnormalized sEMG in control algorithm development for level walking.

Factor Analysis: Dealing with Response Bias

This paper proposes an innovative method for factor analyzing data that potentially contains individual response bias. Past methods include the use of “ipsative” data, or, related to that, “ipsatized” data. Unfortunately, factor analysis as the main method used for analyzing the dimensionality of data, cannot be applied to ipsative data. In contrast, normalization of data as an alternative method to filter out response bias, is not hampered by the technical statistical issues inherent to applying multivariate techniques to ipsative data. Using high-quality data from a survey in Nepal that makes use of – among others – the High-Performance Organizations (HPO) framework, this paper shows that the traditional approach of directly applying Confirmatory Factor Analysis (CFA) starting from an existing model or theory, is inferior to our approach. Even applying Exploratory Factor Analysis (EFA) to the raw (non-normalized) data before using CFA, is unable to detect the optimal dimensionality, or structure, in the data. A better structure can be obtained by performing EFA on normalized data that corrects for response bias in the raw data. This paper convincingly shows that the newly identified structure is superior to the original structure suggested by the HPO framework. Applying a CFA using the newly detected structure on the raw data, gives excellent goodness-of-fit statistics, with more items retained, and no need of forced methods to improve the model fit. The findings suggest that existing models and questionnaires based on these models, are not necessarily as valid and reliable as empirical studies that make use of traditional analyses seem to suggest. When adopting existing instruments, researchers are advised to critically check the validity and reliability of these instruments – especially those vulnerable to response bias - and to apply the procedures laid out in this paper, in order to enhance the quality of their research, and to inform future researchers who consider using the same instruments or to warn them about the potential shortcomings of these instruments.

Towards On-Site Implementation of Multi-Step Air Pollutant Index Prediction in Malaysia Industrial Area: Comparing the NARX Neural Network and Support Vector Regression

Malaysia has experienced public health issues and economic losses due to air pollution problems. As the air pollution problem keeps increasing over time, studies on air quality prediction are also advancing. The air quality prediction can help reduce air pollution’s damaging impact on public health and economic activities. This study develops and evaluates the Nonlinear Autoregressive Exogenous (NARX) Neural Network and Support Vector Regression (SVR) for multi-step Malaysia’s Air Pollutant Index (API) prediction, focusing on the industrial areas. The performance of NARX and SVR was evaluated on four crucial aspects of on-site implementation: Input pre-processing, parameter selection, practical predictability limit, and robustness. Results show that both predictors exhibit almost comparable performance, in which the SVR slightly outperforms the NARX. The RMSE and R2 values for the SVR are 0.71 and 0.99 in one-step-ahead prediction, gradually changing to 6.43 and 0.68 in 24-step-ahead prediction. Both predictors can also perform multi-step prediction by using the actual (non-normalized) data, hence are simpler to be implemented on-site. Removing several insignificant parameters did not affect the prediction performance, indicating that a uniform model can be used at all air quality monitoring stations in Malaysia’s industrial areas. Nevertheless, SVR shows more resilience towards outliers and is also stable. Based on the trends exhibited by the Malaysia API data, a yearly update is sufficient for SVR due to its strength and stability. In conclusion, this study proposes that the SVR predictor could be implemented at air quality monitoring stations to provide API prediction information at least nine steps in advance.

TARNet: An Efficient and Lightweight Trajectory-Based Air-Writing Recognition Model Using a CNN and LSTM Network

Air-writing is a growing research topic in the field of gesture-based writing systems. This research proposes a unified, lightweight, and general-purpose deep learning algorithm for a trajectory-based air-writing recognition network (TARNet). We combine a convolutional neural network (CNN) with a long short-term memory (LSTM) network. The architecture and applications of CNN and LSTM networks differ. LSTM is good for time series prediction yet time-consuming; on the other hand, CNN is superior in feature generation but comparatively faster. In this network, the CNN and LSTM serve as a feature generator and a recognizer, optimizing the time and accuracy, respectively. The TARNet utilizes 1-dimensional separable convolution in the first part to obtain local contextual features from low-level data (trajectories). The second part employs the recurrent algorithm to acquire the dependency of high-level output. Four publicly available air-writing digit (RealSense trajectory digit), character (RealSense trajectory character), smart-band, and Abas datasets were employed to verify the accuracy. Both the normalized and nonnormalized conditions were considered. The use of normalized data required longer training times but provided better accuracy. The test time was the same as those for nonnormalized data. The accuracy for RTD, RTC, smart-band, and Abas datasets were 99.63%, 98.74%, 95.62%, and 99.92%, respectively.

Age-related changes in dynamic corneal backscatter observed in Scheimpflug imaging.

Corneal visualization Scheimpflug technology (Corvis ST) has the potential to indirectly provide information on corneal structure via the analysis of statistical properties of the backscatter. The aim of this work was to ascertain whether there are age-related changes in the dynamics of corneal backscatter during an air-puffed induced corneal deformation. Retrospective data from Corvis ST measurements of 151 young subjects (19-30 years) and 82 older subjects (50-87 years) were considered. Each measurement consisted of 140 frames (sampling frequency: 4330 fps). For every frame the cornea was first segmented, then regions of interest, encompassing temporal, central and nasal parts of cornea were selected, to which the parameters of Weibull distribution (scale and shape) were fitted, leading to time series of the estimated parameters. Apparent differences were found between the parameters of Weibull distribution between the two considered groups that manifest themselves mostly in the nasal region of the cornea. However, those differences cannot be attributed to the age alone. For this, a normalization method is proposed that leads to a much better separation between the groups in all considered regions. Clinical Relevance- The parameters of the corneal backscat-ter are widely used to asses corneal clarity (so-called corneal densitometry). Recently the parameters of Weibull distribution fitted to the corneal backscatter data have been used to support diagnosis of keratoconus. This work contributes to the assessment of corneal clarity by identifying the apparent age-related differences in those parameters when dynamic raw data is considered highlighting the need for such parameters to be appropriately normalized. Further it shown that the shape parameter of Weibull distribution unlike the scale parameter carries that information already for the raw non-normalized data.

Five-week warning of COVID-19 peaks prior to the Omicron surge in Detroit, Michigan using wastewater surveillance

Wastewater-based epidemiology (WBE) is useful in predicting temporal fluctuations of COVID-19 incidence in communities and providing early warnings of pending outbreaks. To investigate the relationship between SARS-CoV-2 concentrations in wastewater and COVID-19 incidence in communities, a 12-month study between September 1, 2020, and August 31, 2021, prior to the Omicron surge, was conducted. 407 untreated wastewater samples were collected from the Great Lakes Water Authority (GLWA) in southeastern Michigan. N1 and N2 genes of SARS-CoV-2 were quantified using RT-ddPCR. Daily confirmed COVID-19 cases for the City of Detroit, and Wayne, Macomb, Oakland counties between September 1, 2020, and October 4, 2021, were collected from a public data source. The total concentrations of N1 and N2 genes ranged from 714.85 to 7145.98 gc/L and 820.47 to 6219.05 gc/L, respectively, which were strongly correlated with the 7-day moving average of total daily COVID-19 cases in the associated areas, after 5 weeks of the viral measurement. The results indicate a potential 5-week lag time of wastewater surveillance preceding COVID-19 incidence for the Detroit metropolitan area. Four statistical models were established to analyze the relationship between SARS-CoV-2 concentrations in wastewater and COVID-19 incidence in the study areas. Under a 5-week lag time scenario with both N1 and N2 genes, the autoregression model with seasonal patterns and vector autoregression model were more effective in predicting COVID-19 cases during the study period. To investigate the impact of flow parameters on the correlation, the original N1 and N2 gene concentrations were normalized by wastewater flow parameters. The statistical results indicated the optimum models were consistent for both normalized and non-normalized data. In addition, we discussed parameters that explain the observed lag time. Furthermore, we evaluated the impact of the omicron surge that followed, and the impact of different sampling methods on the estimation of lag time.

Interindividual variability of human thermoregulation: Toward personalized ergonomics of the indoor thermal environment.

ObjectiveThe study was undertaken to show the magnitude of interindividual differences in energy expenditure (i.e., heat production) under normal living conditions with the aim of providing physiological evidence to support the advancement of a personalized thermal conditioning approach.MethodsThree sets of experimental protocols with six participants were conducted at neutral and mild cold temperatures. Energy expenditure, local skin temperatures, and core body temperature were measured continuously, while cognitive performance and thermal sensation were surveyed intermittently. The protocols were designed to study the effects of several normal day activities, low‐level physical activity and eating a meal, on metabolic and physiological parameters.ResultsLarge interindividual differences among the subjects were demonstrated using non‐normalized data by design. The resting metabolic rate difference was 58%, the percentage change in energy expenditure during standing compared to sitting was up to 31%, and the difference in mechanical work efficiency between the least and the most efficient individual was 39.1%. Energy expenditure increase due to the meal effect was 11.2% to 23.3% at neutral and 9.9% to 33.9% at mild cold temperatures across individuals.ConclusionsLarge interindividual differences in metabolic rate under typical everyday living and office activities suggest facilitating personalized thermal conditioning instead of providing uniform temperature. Therefore, it is necessary to find noninvasive markers that can be easily measured and used as surrogates for human heat production to individualize the climate control of buildings.

Novel and Effective: ECG Utilization for MI Confirmation in Mouse Models

IntroductionElectrocardiography measures the electrical activity of the heart to generate an electrocardiogram (ECG) through a non‐invasive process that involves placing electrodes on strategic points on the body. When an MI occurs, the lack of contractility induces changes to the ECG, visualized by a prolonged QRS wave and amplification of the R wave. Previously, we identified the area under the QRS wave, 1 min after permanent ligation in a mouse model of myocardial infarction (MI) determined surgical success. The purpose of this study was to further demonstrate in a second cohort that the previous ECG parameters identified accurately confirmed that an MI has been administered by left anterior coronary artery ligation in real‐time.Methods and ResultsPermanent coronary artery ligation was performed on WT and CD8atmak1 (CD8‐/‐) mice (n=9/genotype); an ECG was recorded using a surgical monitoring platform and was analyzed on iWorkx LabScribe Data Acquisition and Analysis software. Previous analysis of the first dataset indicated no differences between sexes so both males and females were used in the second cohort. A receiver operating characteristic (ROC) curve was used to link changes in the QRS complex at the 1‐min mark with MI surgical success. By ROC analysis, the area under the QRS complex had an area under the curve (AUC) of 0.94 (p=0.01) indicating a strong association with MI success. To remove any potential effect of surgical platform variability, we normalized the 1 min data to pre‐MI values to determine fold‐change in the area under the QRS complex. ROC analysis of the normalized data also indicated a strong sensitivity and specificity for MI success with an AUC value of 0.90 (p=0.02). Mice in the second cohort were divided into two groups based on a cut‐off point for either the area under the QRS complex (≥0.003 mV/sec) or the normalized data (≥ 3.1 fold change in the QRS complex) to distinguish the two groups. Using the non‐normalized data alone, there was an 84% success rate with 2 false positives and 1 false negative. Analysis using the normalized data alone had an 74% success rate, with 2 false positives and 3 false negatives. None of these false positives or negatives were linked to genotype or sex.ConclusionOur data indicates that while changes in the QRS complex 1‐min after ligation may be useful for confirming MI surgical success, there are potential limitations that could skew the data. One possible explanation for the false positives could be variation between ECG boards. To remove this, we normalized post‐MI data to pre‐MI values but, this seemed to decrease effectiveness. It is also important to note that many of the false negatives were due to anomalies within the ECG such as the presence of R‐prime waves. For this reason, users must be consistent when analyzing the ECG data points to minimize variation. In summary, using the ECG data 1 min after ligation had a fairly strong prediction rate for MI surgical success.

Investigating the Sources of Urban Air Pollution Using Low-Cost Air Quality Sensors at an Urban Atlanta Site.

Advances in low-cost sensors (LCS) for monitoring air quality have opened new opportunities to characterize air quality in finer spatial and temporal resolutions. In this study, we deployed LCS that measure both gas (CO, NO, NO2, and O3) and particle concentrations and co-located research-grade instruments in Atlanta, GA, to investigate the capability of LCS in resolving air pollutant sources using non-negative matrix factorization (NMF) in a moderately polluted urban area. We provide a comparison of applying the NMF technique to both normalized and non-normalized data sets. We identify four factors with different temporal trends and properties for both normalized and non-normalized data sets. Both normalized and non-normalized LCS data sets can resolve primary organic aerosol (POA) factors identified from research-grade instruments. However, applying normalization provides factors with more diverse compositions and can resolve secondary organic aerosol (SOA). Results from this study demonstrate that LCS not only can be used to provide basic mass concentration information but also can be used for in-depth source apportionment studies even in an urban setting with complex pollution mixtures and relatively low aerosol loadings.

A comparative simulation of normalization methods for machine learning-based intrusion detection systems using KDD Cup’99 dataset

Network Intrusion detection systems (NIDS) detect malicious and intrusive information in computer networks. Presently, commercial NIDS is based on machine learning approaches that have complex algorithms and increase intrusion detection efficiency and efficacy. These machine learning-based NIDS use high dimensional network traffic data from which intrusive information is to be detected. This high-dimensional network traffic data in NIDS needs to be preprocessed and normalized to make it suitable for machine learning tools. A machine learning approach with appropriate normalization and prepossessing increases NIDS performance. This paper presents an empirical study on various normalization methods implemented on a benchmark network traffic dataset, KDD Cup’99, that has been used to evaluate the NIDS model. The present study shows decimal normalization has a better prediction performance than non-normalized traffic data categorized into ‘normal’ or ‘intrusive’ classes.

Predicting of acid red 14 removals from synthetic wastewater in the advanced oxidation process using artificial neural networks and fuzzy regression

The titanium dioxide (TiO2) photocatalyst reactor is a widely used approach for eliminating non-biodegradable compounds in water and wastewater. Predicting the performance of the reactor in removing acid red 14 (AR14) was investigated using radial basis function (RBF), Adaptive Neuro-Fuzzy Inference System (ANFIS), and fuzzy regression analysis. The input variables in the models were radiation, irradiation time, the initial concentration of nanoparticle TiO2 to cement ratio (nTiO2/c), pH, initial concentration of AR14, flow rate, and initial concentration of oxidizing peroxydisulfate. In addition, for reducing the overfitting of the training process, we used K-fold cross-validation. The results of using the three methods indicated that the ANFIS performance was more suitable in comparison with the fuzzy regression models, such that the coefficient of determination (R2), index of agreement (IA), Nash–Sutcliffe efficiency (E) for model efficiency, Mean Squared Error (MSE), and Mean Bias Error (MBE) for training between the observed data and predicted data reached 0.965, 0.991, 0.73, 0.0132 and 0.019, respectively. Normalized input data improved slightly the training processes and performance of RBF compared with non-normalized input data. The ANFIS neural network for training process and model performance slightly achieved better results than the RBF neural network. The performance of the fuzzy regression approach was less suitable in comparison with RBF and ANFIS modeling. The sensitivity analysis of the RBF neural network depicted that the initial concentration of peroxydisulfate (S2O8) was the most effective factor (25.6%) in predicting the removal of AR14 from synthetic wastewater. After that, the effective factors were irradiation time (12.44%), initial concentration of AR14 (10.37%), pH (10.37), flow rate (8.55%), and the weight ratio of the initial concentration of nanoparticle TiO2 to cement ratio (nTiO2/c) (1.75%), respectively.

Spectral analysis of ultrasound radiofrequency backscatter for the identification of epicardial adipose tissue.

Purpose: The coronary arteries are embedded in a layer of fat known as epicardial adipose tissue (EAT). The EAT influences the development of coronary artery disease (CAD), and increased EAT volume can be indicative of the presence and type of CAD. Identification of EAT using echocardiography is challenging and only sometimes feasible on the free wall of the right ventricle. We investigated the use of spectral analysis of the ultrasound radiofrequency (RF) backscatter for its potential to provide a more complete characterization of the EAT. Approach: Autoregressive (AR) models facilitated analysis of the short-time signals and allowed tuning of the optimal order of the spectral estimation process. The spectra were normalized using a reference phantom and spectral features were computed from both normalized and non-normalized data. The features were used to train random forests for classification of EAT, myocardium, and blood. Results: Using an AR order of 15 with the normalized data, a Monte Carlo cross validation yielded accuracies of 87.9% for EAT, 84.8% for myocardium, and 93.3% for blood in a database of 805 regions-of-interest. Youden's index, the sum of sensitivity, and specificity minus 1 were 0.799, 0.755, and 0.933, respectively. Conclusions: We demonstrated that spectral analysis of the raw RF signals may facilitate identification of the EAT when it may not otherwise be visible in traditional B-mode images.

Analytics on Non-Normalized Data Sources: More Learning, Rather Than More Cleaning

Data analysis is increasingly performed over data assembled from uncontrolled sources, facing inconsistency in knowledge-representation conventions. The typical practice is to create “clean” data for analysis, matching entities and merging variants to overcome differences in knowledge representation. Despite progress in data management techniques to automate this process, it still needs labor-intensive supervision from the analyst. In this paper, we evaluate the benefit of advanced statistical tools to address directly many analytic tasks across data sources without such entity-matching cleaning. Reframing analytical questions as machine-learning tasks enables to replace exact matching of entities by continuous descriptions–vectorial embeddings– that expose similarities between entries. But are analyses with less cleaning trustworthy? We answer this question with a thorough benchmark on questions typical of socio-economic studies across 14 employee databases: we compare the approaches based on machine learning to manual data cleaning (entity matching). It reveals that using embeddings and machine learning improves results validity (smaller estimation error) more than manual cleaning, with considerably less human labor. While machine learning is often combined with data management for the purpose of cleaning, our study suggests that using it directly for <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">analysis</i> is beneficial because it captures ambiguities hard to represent during curation.