Missing Data Research Articles

A probabilistic framework for identifying anomalies in urban air quality data.

Just as the value of crude oil is unlocked through refining, the true potential of air quality data is realized through systematic processing, analysis, and application. This refined data is critical for making informed decisions that may protect health and the environment. Perhaps ground-based air quality monitoring data often face quality control issues, notably outliers. The outliers in air quality data are reported as error and event-based. The error-based outliers are due to instrument failure, self-calibration, sensor drift over time, and the event based focused on the sudden change in meteorological conditions. The event-based outliers are meaningful while error-based outliers are noise that needs to be eliminated and replaced post-detection. In this study, we address error-based outlier detection in air quality data, particularly targeting particulate pollutants (PM2.5 and PM10) across various monitoring sites in Delhi. Our research specifically examines data from sites with less than 5% missing values and identifies four distinct types of error-based outliers: extreme values due to measurement errors, consecutive constant readings and low variance due to instrument malfunction, periodic outliers from self-calibration exceptions, and anomalies in the PM2.5/PM10 ratio indicative of issues with the instruments' dryer unit. We developed a robust methodology for outlier detection by fitting a non-linear filter to the data, calculating residuals between observed and predicted values, and then assessing these residuals using a standardized Z-score to determine their probability. Outliers are flagged based on a probability threshold established through sensitivity testing. This approach helps distinguish normal data points from suspicious ones, ensuring the refined quality of data necessary for accurate air quality modeling. This method is essential for improving the reliability of statistical and machine learning models that depend on high-quality environmental data.

What happens between first symptoms and first acute exacerbation of COPD - observational study of routine data and patient survey.

Chronic obstructive pulmonary disease affects nearly 400 million worldwide - over a million in the United Kingdom - and is the third leading cause of death. However, there is limited understanding of what prompts a diagnosis, how long this takes from symptom onset and the different approaches to clinical management by primary care professionals. Map out the clinical management and National Health Service contacts from symptom presentation to chronic obstructive pulmonary disease diagnosis and first acute exacerbation of chronic obstructive pulmonary disease in three time periods; construct risk prediction for first acute exacerbation of chronic obstructive pulmonary disease. Retrospective cohort study and cross-sectional survey. Primary care. Patients with incident chronic obstructive pulmonary disease aged > 35 years in England. None. First acute exacerbation of chronic obstructive pulmonary disease. Clinical Practice Research Datalink Aurum; new online survey. Forty thousand five hundred and seventy-seven patients were diagnosed between April 2006 and March 2007 (cohort 1), 48,249 between April 2016 and March 2017 (cohort 2) and 4752 between March and August 2020 (cohort 3). The mean (standard deviation) age was 68.3 years (12.0); 47.3% were female. Around three-quarters were diagnosed in primary care, with a slight fall in cohort 3. Compliance with National Institute for Health and Care Excellence diagnostic guidelines was slightly higher in cohorts 2 and 3 for all patients; 35.8% (10.0% in the year before diagnosis) had all four elements met for all cohorts combined. Multilevel modelling showed considerable between-practice variation in spirometry. The survey on the charity website had 156 responses by chronic obstructive pulmonary disease patients. Many respondents had not heard of the condition, hoped the symptoms would go away and identified various healthcare-related barriers to earlier diagnosis. Clinical Practice Research Datalink analysis showed notable changes in post-diagnosis prescribing from cohort 1 to 2, such as increases in long-acting muscarinic antagonist (21.7-46.3%). Triple therapy rose from 2.9% in cohort 2 to 11.1% in cohort 3. Documented pulmonary rehabilitation rose from just 0.8% in cohort 1 to 13.7% in cohort 2 and 20.9% in cohort 3. For all patients combined, the median time to first acute exacerbation of chronic obstructive pulmonary disease in patients who had one was 1.4 years in cohorts 1 and 2. Acute exacerbation of chronic obstructive pulmonary disease prediction models identified some consistent predictors, such as age, deprivation, severity, comorbidities, post-diagnosis spirometry and annual review. Models without post-diagnosis general practitioner actions had a c-statistic of around 0.70; the highest c-statistic was 0.81, for cohort 2 with post-diagnosis general practitioner actions and 6-month follow-up. All models had good calibration. The three most important predictors in terms of their population attributable risks were being a current smoker and offered smoking cessation advice (32.8%), disease severity (30.6%) and deprivation (15.4%). The highest population attributable risks for variables with adjusted hazard ratios < 1 were chronic obstructive pulmonary disease review (-27.3%) and flu vaccination (-26.6%). Symptom recording and chronic obstructive pulmonary disease diagnosis vary between practice; predicted forced expiratory volume in 1 second had many missing values. There has been some improvement over time in chronic obstructive pulmonary disease diagnosis and management, with large changes in prescribing, though patient and system barriers to further improvement exist. Data available to general practitioners cannot generate risk prediction models with sufficient accuracy. It will be important to expand the COVID-era cohort with longer follow-up and augment general practitioner data for better prediction. This study is registered as Researchregistry.com: researchregistry4762. This award was funded by the National Institute for Health and Care Research (NIHR) Health and Social Care Delivery Research programme (NIHR award ref: 17/99/72) and is published in full in Health and Social Care Delivery Research; Vol. 12, No. 43. See the NIHR Funding and Awards website for further award information.

Temporal Autoregressive Matrix Factorization for High-Dimensional Time Series Prediction of OSS.

Open-source software (OSS) plays an increasingly significant role in modern software development tendency, so accurate prediction of the future development of OSS has become an essential topic. The behavioral data of different open-source software are closely related to their development prospects. However, most of these behavioral data are typical high-dimensional time series data streams with noise and missing values. Hence, accurate prediction on such cluttered data requires the model to be highly scalable, which is not a property of traditional time series prediction models. To this end, we propose a temporal autoregressive matrix factorization (TAMF) framework that supports data-driven temporal learning and prediction. Specifically, we first construct a trend and period autoregressive model to extract trend and period features from OSS behavioral data, and then combine the regression model with a graph-based matrix factorization (MF) to complete the missing values by exploiting the correlations among the time series data. Finally, use the trained regression model to make predictions on the target data. This scheme ensures that TAMF can be applied to different types of high-dimensional time series data and thus has high versatility. We selected ten real developer behavior data from GitHub for case analysis. The experimental results show that TAMF has good scalability and prediction accuracy.

Preinjury and Event-Related Characteristics of Pediatric Firearm Injuries: The American College of Surgeons Firearm Study, United States, March 2021‒February 2022.

Objectives. To assess differences in contextual factors by intent among pediatric firearm injury patients and determine factors associated with data missingness. Methods. We retrospectively queried the American College of Surgeons Firearm Study database (March 1, 2021-February 28, 2022) for patients aged 18 years or younger. We stratified preinjury, firearm-related, and event-related factors by intent and compared them by using Fisher exact, χ2, or 1-way analysis of variance testing. Secondary analysis estimated the adjusted odds of missingness by using generalized linear modeling with binominal logit link. Results. Among 17 395 patients, 2974 (17.1%) were aged 18 years or younger; 1966 (66.1%) were injured by assault, 579 (19.5%) unintentionally, and 76 (2.6%) by self-inflicted means. Most contextual factors differed by intent, including proportion of youths with previous adverse childhood experiences, mental illness, and violent assaults or injury, firearm type and access, perpetrator relationship, and injury location. In adjusted analyses, age, trauma center designation, intent, and admission status were associated with missingness. Conclusions. Contextual factors related to pediatric firearm injury vary by intent. Specific predictors associated with missingness may inform improved future data collection. Public Health Implications. Contextual factors related to pediatric firearm injury can be obtained in a systematic manner nationally to inform targeted interventions. (Am J Public Health. 2024;114(10):1097-1109. https://doi.org/10.2105/AJPH.2024.307754).

Bayesian merged utilization of GRAPPA and SENSE (BMUGS) for in-plane accelerated reconstruction increases fMRI detection power

In fMRI, capturing brain activity during a task is dependent on how quickly the k-space arrays for each volume image are obtained. Acquiring the full k-space arrays can take a considerable amount of time. Under-sampling k-space reduces the acquisition time, but results in aliased, or “folded,” images after applying the inverse Fourier transform (IFT). GeneRalized Autocalibrating Partial Parallel Acquisition (GRAPPA) and SENSitivity Encoding (SENSE) are parallel imaging techniques that yield reconstructed images from subsampled arrays of k-space. With GRAPPA operating in the spatial frequency domain and SENSE in image space, these techniques have been separate but can be merged to reconstruct the subsampled k-space arrays more accurately. Here, we propose a Bayesian approach to this merged model where prior distributions for the unknown parameters are assessed from a priori k-space arrays. The prior information is utilized to estimate the missing spatial frequency values, unalias the voxel values from the posterior distribution, and reconstruct into full field-of-view images. Our Bayesian technique successfully reconstructed simulated and experimental fMRI time series with no aliasing artifacts while decreasing temporal variation and increasing task detection power.

Classification of Air Pollutant Index on Data with Outliers and Imbalance Class Problem

The problem of air pollution has become a global issue that has received attention from various countries. Jakarta, Indonesia's capital city, is unavoidable from the same problem. This study will use four parameters of substances PM10, SO2, CO, O3, and nitrogen dioxide to categorize Jakarta's air quality (NO2). The data used is daily data taken from the Air Quality Monitoring Station in Jakarta throughout 2020. The methods used include SVM, Random Forest, Logistic Regression, KNN, CART, and Stacking Algorithm. At the data preparation stage, we found missing values, outliers, and class imbalance problems. Before applying machine learning methods and evaluating accuracy, we used data pre-processing techniques such as the MissForest method, median substitution, and ADASYN. The results prove that the original dataset has a higher accuracy score (0.882 – 0.977) than the balanced dataset (0.829 – 0.976). According to the evaluation results, the Random Forest method has the highest accuracy score for original and balanced datasets. The overall result is better than the identical research, which produces 96.61% accuracy using a neural network. It shows that preprocessing steps such as missing values handling an imbalanced class handling is essential in classification performance.

Predicting the Optimal Treatment for Diseases Using Whale Optimization Algorithm

Recently, there has been a significant research focus on addressing disease prevalence, especially in dealing with the complexities of big data associated with disease data. The challenge is achieving high accuracy due to missing value problems in big data. This study aims to use AI techniques to develop a system that predicates the optimal solutions for disease, regardless of the type of disease, i.e. the system can be applied to any type of disease. The approach involves handling missing values and normalizing disease datasets. The Whale Optimization Algorithm (WOA) will be used to improve predictions for effective disease treatments. We obtained good results for predicting the appropriate treatment for the disease in the proposed research, compared to the results obtained when applying the PSO algorithm before development in state of earlier, where the results obtained in the proposed research had higher accuracy than the results in in state of earlier at high iterations starting from 200 iterations and also had a lower error rate.:

청소년이 인식하는 혐오표현에 관한 연구

The purpose of this study is to explore the factors influencing hate speech during adolescence and to examine the impact of hate awareness formed during this period on various aspects of adolescent life. For this purpose, publicly available data from the “2020 Study on the Exposure to and Countermeasures for Hate Speech among Adolescents,” conducted by the National Youth Policy Institute of Korea, was used. The analysis included data from 3,938 middle and high school students, excluding missing values. The main research findings based on the analysis are as follows. First, it was found that as experiences of hate speech victimization, the perceived social influence of hate speech, and levels of hate towards minority groups increased, the awareness of hate speech during adolescence also heightened. Second, among the risk factors influencing the awareness of hate speech during adolescence, the social influence of hate speech emerged as the most significant factor. Through this, the aim is to raise awareness among adolescents about social minorities, such as women, individuals with disabilities, and the elderly, emphasizing that they are members of our societal community with whom we must coexist.

A New Robust Lunar Landing Selection Method Using the Bayesian Optimization of Extreme Gradient Boosting Model (BO-XGBoost)

The safety of lunar landing sites directly impacts the success of lunar exploration missions. This study develops a data-driven predictive model based on machine learning, focusing on engineering safety to assess the suitability of lunar landing sites and provide insights into key factors and feature representations. Six critical engineering factors were selected as constraints for evaluation: slope, elevation, roughness, hillshade, optical maturity, and rock abundance. The XGBoost model was employed to simulate and predict the characteristics of landing areas and Bayesian optimization was used to fine-tune the model’s key hyperparameters, enhancing its predictive performance. The results demonstrate that this method effectively extracts relevant features from multi-source remote sensing data and quantifies the suitability of landing zones, achieving an accuracy of 96% in identifying landing sites (at a resolution of 0.1° × 0.1°), with AUC values exceeding 95%. Notably, slope was recognized as the most critical factor affecting safety. Compared to assessment processes based on Convolutional Neural Networks (CNNs) and Random Forest (RF) models, XGBoost showed superior performance in handling missing values and evaluating feature importance accuracy. The findings suggest that the BO-XGBoost model shows notable classification performance in evaluating the suitability of lunar landing sites, which may provide valuable support for future landing missions and contribute to optimizing lunar exploration efforts.

An automated approach to predict diabetic patients using KNN imputation and effective data mining techniques

Diabetes is thought to be the most common illness in underdeveloped nations. Early detection and competent medical care are crucial steps in reducing the effects of diabetes. Examining the signs associated with diabetes is one of the most effective ways to identify the condition. The problem of missing data is not very well investigated in existing works. In addition, existing studies on diabetes detection lack accuracy and robustness. The available datasets frequently contain missing information for the automated detection of diabetes, which might negatively impact machine learning model performance. This work suggests an automated diabetes prediction method that achieves high accuracy and effectively manages missing variables in order to address this problem. The proposed strategy employs a stacked ensemble voting classifier model with three machine learning models. and a KNN Imputer to handle missing values. Using the KNN imputer, the suggested model performs exceptionally well, with accuracy, precision, recall, F1 score, and MCC of 98.59%, 99.26%, 99.75%, 99.45%, and 99.24%, respectively. In two scenarios one with missing values eliminated and the other with KNN imputer, the study thoroughly compared the suggested model with seven other machine learning techniques. The outcomes demonstrate the superiority of the suggested model over current state-of-the-art methods and confirm its efficacy. This work demonstrates the capability of KNN imputer and looks at the problem of missing values for diabetes detection. Medical professionals can utilize the results to improve care for diabetes patients and discover problems early.

Spatial interpolation of global DEM using federated deep learning

The digital elevation model (DEM) provides important data support for 3D terrain modeling. However, due to the complex and changeable terrain in the real world and the high cost of field measurement, it is extremely difficult to obtain continuous and high-density elevation data directly. Therefore, it is necessary to rely on spatial interpolation technology to restore the DEM overall picture in the original sampling area. The traditional spatial interpolation method usually has the characteristics of low model complexity and high computational cost, which leads to low real-time performance and low precision of the interpolation process. The interpolation operation based on DEM data can be considered as a special image generation process where the input is a DEM image with missing values and the output is a complete DEM image. At present, a large number of studies have proved that deep learning methods are very effective in image generation tasks. However, the training of deep learning models requires the support of a large number of high-quality data sets. DEM data in various countries, especially in key regions, are usually restricted by privacy protection regulations and cannot be disclosed. The emergence of Federated Learning (FL) provides a new solution, which supports local training on multiple end nodes, without sending local data to a remote center server for centralized training, effectively protecting data privacy. In this study, we propose a DEM interpolation model based on FL and multiScale U-Net. The experimental results show that compared with the traditional method, this model has faster processing speed and lower interpolation precision. At the same time, this research result provides a new way for efficient and secure use of terrain information, especially in those application scenarios that have strict requirements for DEM data privacy and security.

Gender-based differences in the association of self-reported sleep duration with cardiovascular disease and diabetes

ObjectiveInsufficient or prolonged sleep each day may contribute to the onset of cardiovascular disease and diabetes, and there may be some variability between genders; however, current research evidence is limited. We aimed to investigate the effects of gender on self-reported sleep duration and the prevalence of cardiovascular disease and diabetes.Research design and methodsThis study is a population-based, cross-sectional analysis. Data from a nationally representative sample of US adults obtained from the National Health and Nutrition Examination Survey (NHANES) (2005–2020), and 13,002 participants, including 6,774men and 6,228women, were obtained by excluding the missing values for each variable self-reported sleep duration data obtained by using a habitual baseline questionnaire. Logistic regression models investigated the associations between gender-specific self-reported sleep duration, CVDs, and diabetes events.ResultIn all participants, respectively, compared with sleep 7–8 h/day, the multivariable-adjusted odds ratios significantly associated with < 7 h /day and > 8 h /day were (1.43[1.15, 1.78]) and (1.34[1.01, 1.76]) for CHF, (1.62[1.28, 2.06]) for Angina, (1.42[1.17, 1.71]) for heart attack, (1.38[1.13, 1.70]) and (1.54[1.20, 1.97]) for Stroke, (1.21[1.09, 1.35]) and (1.28[1.11, 1.48]) for diabetes. In men, CHF (1.67[1.21, 2.14]), Angina (1.66[1.18, 2.15]), Stroke (1.55[1.13,1.97]), and diabetes (1.15[1.00, 1.32]) were significantly associated with < 7 h /day, and stroke (1.73[1.16, 2.32]) and diabetes (1.32[1.06, 1.52]) were significantly associated with > 8 h /day. In women, angina(1.83[1.16, 2.50]), heart attack(1.63[1.11, 2.15]), and diabetes (1.32[1.11, 1.54]) were significantly associated with < 7 h /day, while diabetes (1.31[1.03, 1.59]) was significantly associated with > 8 h /day.ConclusionSelf-reported long and short sleep duration was independently associated with partial CVDs and diabetes risk. However, sleep duration and gender did not have multiplicative or additive interactions with the onset of diabetes and CVDs.

Disparities in the diagnostic efficacy of radiomics models in predicting various degrees of cognitive impairment in patients with cerebral small vessel disease

BackgroundAim to validate the diagnostic efficacy of radiomics models for predicting various degrees of cognitive impairment in patients with cerebral small vessel disease (CSVD).MethodsParticipants were divided into mild cognitive impairment group (mild-CSVD group) and sever cognitive impairment group (sever-CSVD group) according to Montreal Cognitive Assessment (MoCA) performance, 98 gender-age-education matched subjects served as normal controls. Radiomic features were extracted from the segmented hippocampus using PyRadiomics. The feature preprocessing involved replacing missing values with the mean, applying stratified random sampling to allocate subjects into training (80%) and testing (20%) sets, ensuring balance among the three classes (normal controls, mild-CSVD group, and sever-CSVD group). A feature selection method was applied to identify discriminative radiomic features, with the optimal texture feature chosen for developing diagnostic models. Performance was evaluated in both the training and testing sets using receiver operating characteristic (ROC) curve analysis.ResultsThe radiomics model achieved an accuracy of 0.625, an AUC of 0.593, a sensitivity of 0.828, and a specificity of 0.316 in distinguishing mild-CSVD group from normal controls. When distinguishing mild-CSVD group from sever-CSVD group, the radiomics model reached an accuracy of 0.683, an AUC of 0.660, a sensitivity of 0.167, and a specificity of 0.897. Similarly, in distinguishing sever-CSVD group from normal controls, the radiomics model exhibited an accuracy of 0.781, an AUC of 0.818, a sensitivity of 0.538, and a specificity of 0.947.ConclusionRadiomics model based on hippocampal texture had disparities in the diagnostic efficacy of radiomics models in predicting various degrees of cognitive impairment in patients with CSVD.

Validation of the Mediating Effect of Parenting Stress in the Relationship between Work-Family Strains and Gains and Marital Satisfaction for Fathers in Korea.

This study aimed to examine the relationship between Korean fathers' work-family strains and gains and marital satisfaction and to validate the mediating effect of parenting stress in the relationship. This study used the 14th-year survey data of the Panel Study of Korean Children (PSKC) conducted by the Korea Institute of Child Care and Education (KICCE), the latest available data. The PSKC is a longitudinal study of children's growth and development from birth in the Republic of Korea. After excluding datasets containing missing values from the data collected from 2150 respondents, the data were extracted from the survey responses collected from 1130 respondents, and descriptive statistical analysis, path analysis, and mediating effect analysis were conducted. First, Korean fathers' work-family strains had a positive (+) influence on parenting stress (β = 0.658, p < 0.001). Second, Korean fathers' work-family strains had a negative (-) influence on marital satisfaction (β = -0.185, p < 0.001). Third, Korean fathers' work-family gains did not have a statistically significant influence on parenting stress (β = 0.016, p = 0.632). Fourth, Korean fathers' work-family gains had a positive (+) influence on marital satisfaction (β = 0.381, p < 0.001). Fifth, Korean fathers' parenting stress did not have a statistically significant influence on marital satisfaction (β = 0.011, p = 0.783). Sixth, the indirect effect of work-family strains on marital satisfaction via parenting stress (β = 0.013, 95% Bias-corrected CI = -0.042∼0.075) was statistically significant, and specifically, it had a partial mediating effect. Seventh, the indirect effect of work-family gains on marital satisfaction via parenting stress (β = 0.028, 95% Bias-corrected CI = 0.005∼0.051) was statistically significant, and specifically, it had a partial mediating effect. This study aims to provide the basic data for various family support policies for fathers by understanding the current status of work-family balance for fathers in modern society, as well as by examining the effects of parenting stress in the relationship with marital satisfaction.

Comparative Analysis of Imputation Methods for Enhancing Predictive Accuracy in Data Models

The presence of missing values within datasets can introduce a detrimental bias, significantly impeding the predictive algorithm's ability to discern patterns and accurately execute prediction. This paper aims to elucidate the intricacies of data imputation methods, providing a more profound understanding of prevalent imputation methods, including list-wise deletion (IGN), mean imputation (AVG), K-Nearest Neighbors (KNN), MissForest (MF), and Predictive Mean Matching (PMM). The dataset employed in this study consists of financial data about S&amp;P 500 companies in the Compustat North America database. The training and validation dataset encompasses 1973 instances, consisting of data during the fourth quarter of 2009, the first quarter of 2010, and the third quarter of 2014. Within this set, 457 missing values were identified and imputed. The test dataset comprises 197 randomly selected instances from the fourth quarter of 2014, equivalent to ten percent of the total instances in the training dataset. The evaluation findings prominently position the dataset derived from MF imputation as the leading performer among all the imputed datasets. The insights derived from this study are intended to assist practitioners in making informed choices when selecting the most suitable data imputation method, particularly in the context of predictive modeling tasks.

Implementing Stacking with Cross - Validation for Heart Disease Prediction.

The objective of this work is to apply machine learning techniques for the prediction and early identification of cardiovascular disease, a major worldwide health problem. XGBoost, K-nearest neighbors (KNN), decision tree (DT), support vector machine (SVM), and StackingCVClassifier were among the ensemble algorithms used to anticipate cardiac disease using a dataset that came from the UCI ML database. To improve the quality of the data, an exploratory data analysis was performed on the dataset using methods including outlier identification and missing value imputation. Stacking CV Classifier attained the best accuracy rate of 92.20%, according to a comparative examination of pre-processing and post-processing findings. When compared to previous approaches, the suggested strategies performed better in terms of accuracy, recall, and f1-score. Moreover, the flexibility of the model indicates its possible application to other illnesses with comparable features.

Reconstructing daytime and nighttime MODIS land surface temperature in desert areas using multi-channel singular spectrum analysis

The availability of continuous spatiotemporal land surface temperature (LST) with high resolution is critical for many disciplines including hydrology, meteorology, ecology, and geology. Like other remote sensing data, satellite–based LST is also encountered with the cloud issue. In this research, over 5000 daytime and nighttime MODIS–LST images are utilized during 2014–2020 for Yazd–Ardakan plain in Yazd, Iran. The multi–channel singular spectrum analysis (MSSA) model is employed to reconstruct missing values due to dusts, clouds, and sensor defect. The selection of eigenvalues is based on the Monte Carlo test and the spectral analysis of eigenvalues. It is found that enlarging the window size has no effect on the number of significant components of the signal which account for the most variance of the data. However, data variance changes for all the three components. Employing two images per day, window sizes 60, 180, 360, and 720 are examined for reconstructing one year LST, where these selections are based on monthly, seasonal, semi-annual, and annual LST cycles, respectively. The results show that window size 60 had the least computational cost and the highest accuracy with RMSE (root mean square error) of 2.6 °C for the entire study region and 1.4 °C for a selected pixel. The gap–filling performance of MSSA is also compared with the one by the harmonic analysis of time series (HANTS) model, showing the superiority of MSSA with an improved RMSE of about 2.7 °C for the study region. In addition, daytime and nighttime LST series for different land covers are compared. Lastly, the maximum, minimum, and average LST for each day and night as well as average and standard deviation of LST images in the seven-year-long time series are also computed.

CMFHMDA: a prediction framework for human disease-microbe associations based on cross-domain matrix factorization.

Predicting associations between microbes and diseases opens up new avenues for developing diagnostic, preventive, and therapeutic strategies. Given that laboratory-based biological tests to verify these associations are often time-consuming and expensive, there is a critical need for innovative computational frameworks to predict new microbe-disease associations. In this work, we introduce a novel prediction algorithm called Predicting Human Disease-Microbe Associations using Cross-Domain Matrix Factorization (CMFHMDA). Initially, we calculate the composite similarity of diseases and the Gaussian interaction profile similarity of microbes. We then apply the Weighted K Nearest Known Neighbors (WKNKN) algorithm to refine the microbe-disease association matrix. Our CMFHMDA model is subsequently developed by integrating the network data of both microbes and diseases to predict potential associations. The key innovations of this method include using the WKNKN algorithm to preprocess missing values in the association matrix and incorporating cross-domain information from microbes and diseases into the CMFHMDA model. To validate CMFHMDA, we employed three different cross-validation techniques to evaluate the model's accuracy. The results indicate that the CMFHMDA model achieved Area Under the Receiver Operating Characteristic Curve scores of 0.9172, 0.8551, and 0.9351$\pm $0.0052 in global Leave-One-Out Cross-Validation (LOOCV), local LOOCV, and five-fold CV, respectively. Furthermore, many predicted associations have been confirmed by published experimental studies, establishing CMFHMDA as an effective tool for predicting potential disease-associated microbes.

A Method to Handle the Missing Values in Multi-Criteria Sorting Problems Based on Dominance Rough Sets

The handling of missing attribute values remains a challenging and problematic issue in data analysis. Imputation techniques are key procedures used to deal with missing attribute values. However, although these methods are widely used, they cause data bias. Rough set theory, a unique mathematical tool for decision making under uncertainty, overcomes this problem by properly adjusting the relationships. Rough sets are often preferred in both classification and sorting problems. The aim of sorting problems is to sort the objects in the decision table (DT) from best to worst and/or to select the best one. For this purpose, it is necessary to obtain a pairwise comparison table (PCT) from the DT. However, in the presence of missing values, the transformation from DT to PCT is not feasible because there are no ranking methods in the literature for sorting problems based on rough sets. To address this limitation, this paper presents a way to transform from DT to PCT and introduces a generalization of the relation belonging to the “do not care” type of missing values in the dominance-based rough set approach (DRSA) to the decision support tool jRank. We also adapted the DomLem algorithm to enable it to work in PCT with missing values. We applied our method step by step to a decision table with 11 objects and investigated the effect of missing values. The experimental results showed that our proposed approach captures the semantics of `do not care’ type missing values.

Abstract B019: A machine learning approach to predicting smoking cessation outcomes among Spanish-speaking smokers who completed a culturally targeted intervention

Abstract Introduction: Among the Hispanic/Latinx population in the US, four of five leading causes of death are attributed to smoking. Few evidence-based interventions have been developed to improve smoking cessation for this population. Our team has demonstrated the efficacy of a culturally targeted, extended self-help intervention for Spanish-speaking smokers, estimating a smoking abstinence rate of 33% at 24 months, compared to 24% for usual care. Further efforts are needed to enhance this efficacious, low-cost, scalable intervention approach. Machine learning (ML) is one approach to increase understanding of the predictors of treatment outcomes and inform strategies to improve the intervention. Toward that goal, this secondary data analysis utilized the decision tree (DT) model to predict self-reported 7-day point prevalence abstinence at 18-month follow-up for participants receiving our intervention. Method: Data from participants who reported smoking status at an 18-month follow-up (N=332 of 714 enrolled, 36% abstinent) were randomly split (80:20) into training and test datasets based on smoking status. We entered demographics, psychosocial, and smoking variables collected at baseline as predictors in the model. In addition to handling missing values and normalizing numeric features, we employed the Synthetic Minority Over-sampling Technique (SMOTE) combined with Tomek links (SMOTE-Tomek) to improve inter-class separability. Recursive Feature Elimination (RFE) using DT was implemented to identify the most relevant predictors. The cross-validation (CV) pipeline incorporated preprocessing, feature selection, handling class-imbalance, over-sampling, and model training with a DT as the interpretable classifier. Hyperparameters were tuned using stratified K-fold (K=5) CV to optimize parameters using Grid search. F1 score was selected as the primary metric as it accounts for each class performance by combining the precision and recall metrics. Finally, the performance of the classifier was evaluated in 20% of the unseen dataset. Results: The DT classifier for the CV achieved the F1 score of 0.73 [95% CI 0.66, 0.77], revealing a reasonably good performance in identifying smokers at 18-months. RFE selected familism, age, affect, and confidence as the most relevant predictors of smoking status. Individuals who had lower familism scores (&amp;lt;138), combined with higher age (&amp;gt;50), were more likely to be smokers at 18 months. Individuals who had higher familism scores (&amp;gt;138) combined with higher negative affect (&amp;gt;42) were more likely to be smokers. Conclusion: This study provides the first step toward personalized care for smoking cessation among the Hispanic/Latinx population and demonstrates the potential of DT classifier to predict cessation outcomes among individuals who received and completed a culturally and linguistically targeted smoking cessation intervention. Our future analysis will compare this model with conventional statistical modeling approaches and other ML algorithms to identify the best-performing parsimonious model. Citation Format: Ranjita Poudel, K. Ruwani M. Fernando, Matthew B. Schabath, Steven K. Sutton, Thomas H. Brandon, Issam El Naqa, Vani N. Simmons. A machine learning approach to predicting smoking cessation outcomes among Spanish-speaking smokers who completed a culturally targeted intervention [abstract]. In: Proceedings of the 17th AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2024 Sep 21-24; Los Angeles, CA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2024;33(9 Suppl):Abstract nr B019.