Nearest-neighbor Imputation Research Articles

Comparison of Imputation Methods for Categorical Real-World Prostate Cancer Data with Natural Order.

Missing values (NA) often occur in cancer research, which may be due to reasons such as data protection, data loss, or missing follow-up data. Such incomplete patient information can have an impact on prediction models and other data analyses. Imputation methods are a tool for dealing with NA. Cancer data is often presented in an ordered categorical form, such as tumour grading and staging, which requires special methods. This work compares mode imputation, k nearest neighbour (knn) imputation, and, in the context of Multiple Imputation by Chained Equations (MICE), logistic regression model with proportional odds (mice_polr) and random forest (mice_rf) on a real-world prostate cancer dataset provided by the Cancer Registry of Rhineland-Palatinate in Germany. Our dataset contains relevant information for the risk classification of patients and the time between date of diagnosis and date of death. For the imputation comparison, we use Rubin's (1974) Missing Completely At Random (MCAR) mechanism to remove 10%, 20%, 30%, and 50% observations. The results are evaluated and ranked based on the accuracy per patient. Mice_rf performs significantly best for each percentage of NA, followed by knn, and mice_polr performs significantly worst. Furthermore, our findings indicate that the accuracy of imputation methods increases with a lower number of categories, a relatively even proportion of patients in the categories, or a majority of patients in a particular category.

Collaborative Filtering for the Imputation of Patient Reported Outcomes.

This study addresses the prevalent issue of missing data in patient-reported outcome datasets, particularly focusing on head and neck cancer patient symptom ratings sourced from the MD Anderson Symptom Inventory. Given that many data mining and machine learning algorithms necessitate complete datasets, the accurate imputation of missing data as an initial step becomes crucial. In this study we propose, for the first time, the use of collaborative filtering for imputing missing head and neck cancer patient symptom ratings. Two configurations of collaborative filtering, namely patient-based and symptom-based, leverage known ratings to infer the missing ones. Additionally, this study compares the performance of collaborative filtering with alternative imputation methods such as Multiple Imputation by Chained Equations, Nearest Neighbor Imputation, and Linear interpolation. Performance is compared using Root Mean Squared Error and Mean Absolute Error metrics. Findings demonstrate that collaborative filtering is a viable and comparatively superior approach for imputing missing patient symptom data.

Clustering polycystic ovary syndrome laboratory results extracted from a large internet forum with machine learning

BackgroundPolycystic Ovary Syndrome (PCOS) is reported to affect between 4% and 21% of reproductive aged people with ovaries. It is a heterogeneous condition with a lack of established phenotypes that address the range of reproductive and metabolic features present in PCOS. These reproductive and metabolic features may result in patients undergoing a variety of relevant laboratory tests. Previous work has led to the gathering of laboratory test results from a PCOS specific forum, hosted on a website called reddit. ObjectivesIn this paper, laboratory results and body mass index (BMI) posted on the PCOS reddit forum were clustered to show the usefulness of the PCOS forum for PCOS research and validate existing PCOS phenotypes or discover other appropriate phenotypes. Methods and resultsOver 1500 sets of PCOS-related reddit laboratory test results and BMIs were clustered using nearest neighbour imputation and K-means clustering. However, only non-imputed data was included in the final clusters. Kernel Density Estimation plots were used to display the distinct clusters. The clustered test results suggested the existence of distinct metabolic and reproductive phenotypes, as well as a group displaying mild features of both types of dysregulations and a group skewed towards normal results. It was also possible to separate the groups further into distinct hypothyroid groups within the mixed dysregulation group and to separate insulin resistant and diabetes-like groups within the metabolic group. ConclusionsThis research further validates the usefulness of exploring alternate data sources in the age of the internet and machine learning. The reddit clusters reinforced the existing notion that people with PCOS can be separated into a primarily metabolic pathology group, a primarily reproductive pathology group and an in between group with pathology in both domains.

A study on the correlation of agricultural carbon emissions in Liaoning Province based on the Spearman correlation coefficient

With the rapid development in the field of modern agriculture, an in-depth investigation into various types of agricultural carbon emissions provides essential insights into understanding the spatial patterns and evolving trends of agricultural carbon emissions in China. Taking Liaoning Province as a case study, this paper collects historical monitoring data on agricultural carbon emissions from all provinces and direct-controlled municipalities in China from 2000 to 2020. The data undergoes preprocessing, with missing values addressed using the nearest-neighbor imputation method. Subsequently, based on the IPCC carbon emission coefficient method, the paper calculates the annual agricultural carbon emissions for various categories in Liaoning Province. The study employs scatter plots to make preliminary judgments on the correlation between different carbon emission categories. Finally, an in-depth analysis is conducted using the Spearman correlation coefficient method to explore the relationships among different carbon sources. The research reveals a high correlation among certain carbon sources, providing scientific guidance for reducing agricultural carbon emissions.

Improving Donor Imputation Using the Prediction Power of Random Forests: a Combination of SwissCheese and missForest

Abstract Imputation procedures are frequently used to treat nonresponse. With random hot deck imputation, missing values are replaced by valid observed values from other units in the same dataset. The recently developed balanced nearest neighbor imputation method, implemented in the SwissCheese R package, generates random hot deck imputation under certain balancing constraints to decrease the variance of the total estimator, in the presence of multivariate nonresponse. The method relies on a notion of neighborhood between units, utilizing a distance measure that becomes difficult to define in high dimensions. In contrast to hot deck imputation methods, many imputation procedures obtain replacement values from prediction models fit from observed data. The missForest method, which uses random forests as prediction models, is an example of this approach. In this article, we propose a new approach that uses the two methods in a complementary manner. We refine the distance measure in the SwissCheese method using missForest predictions. Through a simulation study on empirical data from the Swiss Survey on Income and Living Conditions, we demonstrate reductions in Monte Carlo variance, bias, and mean squared error of the totals obtained by our proposed imputed estimator compared to those obtained using SwissCheese alone.

P264 Machine learning can accurately predict development of Inflammatory Bowel Disease

Abstract Background Prediction of inflammatory bowel disease (IBD) development is the subject of intense study. We have previously shown that significant changes occur in routine laboratory tests in the 5 years preceding IBD diagnosis. Using routinely collected laboratory results, we developed machine learning models to predict development of IBD. Methods We extracted data from the electronic medical records of Maccabi Health Services and included patients with IBD ≥ 16 years of age with a minimum of 5 years follow up and an age and sex matched healthy cohort. The outcome measured was entry into the IBD registry. The study population was split into training and validation cohorts and using laboratory values in the 5 years prior to IBD diagnosis as features, we assessed the performance of supervised learning models and performed interpretability analysis with unsupervised dimensionality reduction. Results 5643 patients with IBD and 17199 healthy people were included in this study. 3039 (53.8%) had Crohn’s disease (CD), 2322 (41.1%) ulcerative colitis (UC) and 282 (5%) indeterminate colitis. The mean age of the IBD and healthy cohorts at study inclusion was 39.3 ± 16.5 and 38.5 ± 15.6 years, respectively. Weighted nearest neighbour imputation was used to correct for data missingness. Unsupervised Uniform Manifold Approximation Projection dimensionality reduction correctly separated the cohort into healthy and IBD clusters (Figure 1A). Furthermore, this algorithm identified two distinct groups correlating to CD or UC diagnosis (Figure 1B). Supervised learning models, namely neural net (AUC 0.95), linear support vector machine (SVM) (AUC 0.93), radial basis function SVM (AUC 0.85) and logistic regression (AUC 0.93) models, accurately predicted IBD diagnosis (Figure 2). These models were repeated without C-reactive protein and the predictive value was not significantly changed. Moreover, we excluded data in the 2 years preceding IBD diagnosis to determine whether model performance was reduced and the predictive value remained the same. Interestingly, using only data from 3-5 years preceding IBD diagnosis, the model could better distinguish between CD and UC (Figure 3). Conclusion In this study, machine learning models could accurately predict a diagnosis of IBD using laboratory data from the 5 years preceding IBD diagnosis. Prospective studies are needed to test these models in a real world setting.

Joint wind and ice hazard for transmission lines in mountainous terrain

This paper focuses on analyzing the ice and wind conditions for a high voltage transmission line system going over mountainous terrain in British Columbia, Canada. After analyzing sixteen surrounding weather stations, an Artificial Neural Network (ANN) is performed for wind speed predictions, the Inverse Distance Weighted Interpolation method (IDW) is used for temperature estimations, and the K Nearest Neighbor Imputation (KNNI) is performed for precipitation rate on the unsampled site of interest. In-cloud icing is estimated using a simple empirical equation proposed by Makkonen & Ahti (1995), in addition to the ice accretion model proposed by Makkonen (2000). The weather data from the described procedure is compared with data from a Weather Research and Forecast model (WRF) reported in previous studies. The wind speed and ice accretion data are fitted to the Weibull and Generalized Pareto distributions, respectively, and five hundred years of data pairs are simulated in addition to the historical data based on those distributions. Finally, hazard contours are drawn, and the results are compared with the combined wind and ice values proposed in the National Standard of Canada (CSA-C22.3) for different reliability levels.

Mapping with height and spectral remote sensing implies that environment and forest structure jointly constrain tree community composition in temperate coniferous forests of eastern Washington, United States

Maps of species composition are important for assessing a wide range of ecosystem functions in forested landscapes, including processes shaping community structure at broader (e.g., climate) and finer (e.g., disturbance) scales. Incorporating recently available remotely sensed datasets has the potential to improve species composition mapping by providing information to help predict species presence and relative abundance. Using USDA Forest Service Forest Inventory and Analysis plot data and the gradient nearest neighbor imputation modeling approach in eastern Washington, USA, we developed tree species composition and structure maps based on climate, topography, and two sources of remote sensing: height from digital aerial photogrammetry (DAP) of pushbroom aerial photography and Sentinel-2 multispectral satellite imagery. We tested the accuracy of these maps based on their capacity to predict species occurrence and proportional basal area for 10 coniferous tree species. In this study region, climate, topography, and location explained much of the species occurrence patterns, while both DAP and Sentinel-2 data were also important in predicting species proportional basal area. Overall accuracies for the best species occurrence model were 68–92% and R2 for the proportional basal area was 0.08–0.55. Comparisons of model accuracy with and without remote sensing indicated that adding some combination of DAP metrics and/or Sentinel-2 imagery increased R2 for the proportional basal area by 0.25–0.45, but had minor and sometimes negative effects on model skill and accuracy for species occurrence. Thus, species ranges appear most strongly constrained by environmental gradients, but abundance depends on forest structure, which is often determined by both environment and disturbance history. For example, proportional basal area responses to moisture limitation and canopy height varied by species, likely contributing to regional patterns of species dominance. However, local-scale examples indicated that remotely sensed forest structures representing recent disturbance patterns likely impacted tree community composition. Overall, our results suggest that characterizing geospatial patterns in tree communities across large landscapes may require not only environmental factors like climate and topography, but also information on forest structure provided by remote sensing.

Stochastic multicriteria acceptability analysis as a forest management priority mapping approach based on airborne laser scanning and field inventory data

The mapping of ecosystem service (ES) provisioning often lacks decision-makers’ preferences on the ESs provided. Analyzing the related uncertainties can be computationally demanding for a landscape tessellated to a large number of spatial units such as pixels. We propose stochastic multicriteria acceptability analyses to incorporate (unknown or only partially known) decision-makers’ preferences into the spatial forest management prioritization in a Scandinavian boreal forest landscape. The potential of the landscape for the management alternatives was quantified by airborne laser scanning based proxies. A nearest-neighbor imputation method was applied to provide each pixel with stochastic acceptabilities on the alternatives based on decision-makers’ preferences sampled from a probability distribution. We showed that this workflow could be used to derive two types of maps for forest use prioritization: one showing the alternative that a decision-maker with given preferences should choose and another showing areas where the suitability of the forest structure suggested different alternative than the preferences. We discuss the potential of the latter approach for mapping management hotspots. The stochastic approach allows estimating the strength of the decision with respect to the uncertainty in both the proxy values and preferences. The nearest neighbor imputation of stochastic acceptabilities is a computationally feasible way to improve decisions based on ES proxy maps by accounting for uncertainties, although the need for such detailed information at the pixel level should be separately assessed.

Pre-Treatment Hemoglobin Concentration and Absolute Monocyte Count as Independent Prognostic Factors for Survival in Localized or Locally Advanced Prostate Cancer Patients Undergoing Radiotherapy.

The prognostic value of inflammatory indices, such as the absolute monocyte count (AMC), has been a subject of interest in recent prostate cancer (PCa) studies, while hemoglobin concentration (HGB) has been recognized as a survival factor in castration-resistant metastatic prostate cancer, but its value remains unclear in localized diseases. The aim of this study was to test the prognostic value of these two simple and inexpensive biomarkers for survival and was based on a cohort of 1016 patients treated with primary radiotherapy and androgen deprivation therapy for localized or locally advanced intermediate- or high-risk PCa. Complete survival data were available for all cases and were based on the National Cancer Registry, with a median observation time of 120 months (Interquartile Range (IQR) 80.9–144.7). Missing blood test data were supplemented using the Nearest Neighbor Imputation, and the Cox Proportional Hazards Regression model was used for analysis. The median age was 68.8 years (IQR 63.3–73.5). The five-year overall survival was 82.8%, and 508 patients were alive at the time of analysis. The median time between blood tests and the first day of radiotherapy was 6 days (IQR 0–19). HGB (p = 0.009) and AMC (p = 0.003) were independent prognostic factors for survival, along with age, Gleason Grade Group, clinical T stage and maximum prostate-specific antigen concentration. This study demonstrates that HGB and AMC can be useful biomarkers for overall survival in patients treated with radiotherapy for localized intermediate- or high-risk PCa.

K Nearest Neighbor Imputation Performance on Missing Value Data Graduate User Satisfaction

A missing value is a common problem of most data processing in scientific research, which results in a lack of accuracy of research results. Several methods have been applied as a missing value solution, such as deleting all data that have a missing value, or replacing missing values with statistical estimates using one calculated value such as, mean, median, min, max, and most frequent methods. Maximum likelihood and expectancy maximization, and machine learning methods such as K Nearest Neighbor (KNN). This research uses KNN Imputation to predict the missing value. The data used is data from a questionnaire survey of graduate user satisfaction levels with seven assessment criteria, namely ethics, expertise in the field of science (main competence), foreign language skills, foreign language skills, use of information technology, communication skills, cooperation, and self-development. The results of testing imputation predictions using KNNI on user satisfaction level data for STMIK PPKIA Tarakanita Rahmawati graduates from 2018 to 2021. Where using the five k closest neighbors, namely 1, 5, 10, 15, and 20, the error value of the k nearest neighbors is 5 in RMSE is 0, 316 while the error value using MAPE is 3,33 %, both values are smaller than the value of k other nearest neighbors. K nearest neighbor 5 is the best imputation prediction result, both calculated by RMSE and MAPE, even in MAPE the error value is below 10%, which means it is very good.

Nearest neighbor ratio imputation with incomplete multinomial outcome in survey sampling.

Nonresponse is a common problem in survey sampling. Appropriate treatment can be challenging, especially when dealing with detailed breakdowns of totals. Often, the nearest neighbor imputation method is used to handle such incomplete multinomial data. In this article, we investigate the nearest neighbor ratio imputation estimator, in which auxiliary variables are used to identify the closest donor and the vector of proportions from the donor is applied to the total of the recipient to implement ratio imputation. To estimate the asymptotic variance, we first treat the nearest neighbor ratio imputation as a special case of predictive matching imputation and apply the linearization method of Yang and Kim (2020). To account for the non-negligible sampling fractions, parametric and generalized additive models are employed to incorporate the smoothness of the imputation estimator, which results in a valid variance estimator. We apply the proposed method to estimate expenditures detail items based on empirical data from the 2018 collection of the Service Annual Survey, conducted by the United States Census Bureau. Our simulation results demonstrate the validity of our proposed estimators and also confirm that the derived variance estimators have good performance even when the sampling fraction is non-negligible.

Hybrid Missing Value Imputation Algorithms Using Fuzzy C-Means and Vaguely Quantified Rough Set

In real cases, missing values tend to contain meaningful information that should be acquired or should be analyzed before the incomplete dataset is used for machine learning tasks. In this work, two algorithms named jointly fuzzy C-Means and vaguely quantified nearest neighbor (VQNN) imputation (JFCM-VQNNI) and jointly fuzzy C-Means and fitted VQNN imputation (JFCM-FVQNNI) have been proposed by considering clustering conception and sufficient extraction of uncertain information. In the proposed JFCM-VQNNI and JFCM-FVQNNI algorithm, the missing value is regarded as a decision feature, and then, the prediction is generated for the objects that contain at least one missing value. Specially, as for JFCM-VQNNI algorithm, indistinguishable matrixes, tolerance relations, and fuzzy membership relations are adopted to identify the potential closest filled values based on corresponding similar objects and related clusters. On the basis of JFCM-VQNNI algorithm, JFCM-FVQNNI algorithm synthetic analyzes the fuzzy membership of the dependent features for instances with each cluster. In order to fill the missing values more accurately, JFCM-FVQNNI algorithm performs fuzzy decision membership adjustment in each object with respect to the related clusters by considering highly relevant decision attributes. The experiments have been carried out on five datasets. Based on the analysis of root-mean-square error, mean absolute error, comparison of imputation values with actual values, and classification accuracy results analysis, we can draw the conclusion that the proposed JFCM-FVQNNI and JFCM-VQNNI algorithms yields sufficient and reasonable imputation performance results by comparing with fuzzy C-Means parameter-based imputation algorithm and fuzzy C-Means rough parameter-based imputation algorithm.

Uncovering forest dynamics using historical forest inventory data and Landsat time series

While many countries regularly produce timely and accurate estimates for a range of forest attributes through national forest inventory programs, many have access to no such inventory or their inventory is discontinuous in space and time, such as a sample-based approach in Ukraine limited to two regional (Ivano-Frankivsk and Sumy) inventories conducted in 2008–2015 at a total area of 37,700 ha. This study addresses the extent to which limited historical forest inventory data may facilitate forest mapping efforts using dense Landsat time series (LTS) and mapping techniques (i.e., classification and imputation) to predict multiyear forest dynamics. We used the Continuous Change Detection and Classification (CCDC) segmentation approach to extract inter- and intra-annual trends in LTS and utilized these fitted trends as covariates in further modeling. We developed random forest (RF) classification models for forest mapping based on field data collection and LTS, and then generated yearly forest maps for 1990–2020. The RF model accuracies were high for both producer’s and user’s accuracy (>0.93 ± 0.04). Based on yearly forest dynamics, we detected an increasing trend in forest loss during 2000–2010 and 2010–2020 in our regions. We used the gradient nearest neighbor (GNN) method to generate species presence (basal area (BA) ≥ 1.0 m2 ha−1) and growing stock volume (GSV) maps at the 30-m pixel level. Our nearest neighbor imputation models achieved better accuracies (Cohen’s kappa > 0.4) for species that are prevalent within regions, occupy distinct geographical areas, and have higher plot BA abundance. To obtain estimates of species abundance with higher R2, we used the GNN model with k = 3 nearest neighbors. Aggregating species based on similar environmental niche resulted in greater R2 values of BA that varied within two regions from low (0.129 and 0.416) for a group of hardwood deciduous species to high (0.531 and 0.712) for coniferous species. The accuracy was systematically better at 5-km hexagonal level, thus we recommend using the maps at coarser aggregations. The study showed that forests accumulated more GSV in 1990–2000, albeit higher rates of GSV loss were observed during the two last decades. We attribute the spatial and temporal character of forest change to forestry activities in the past, the current age structure of forests, and afforestation of abandoned farmlands. We are aware that forecasting performance of the nearest neighbor imputation approach merits more detailed consideration as new forest inventory data are available.

Nearest neighbor imputation for categorical data by weighting of attributes

Missing values are a common phenomenon in modern medical research of complex diseases. The data often contains nominal or categorical variables, for example, single nucleotide polymorphisms (SNPs) in genetic studies. If the missing values are not handled properly, the downstream statistical analysis of incomplete data may be biased. While various imputation methods are available for metrically scaled variables, methods for categorical data are scarce. An imputation method that has been shown to work well for high dimensional metrically scaled variables is the imputation by nearest neighbor methods. In this paper, we propose a weighted nearest neighbors approach to impute missing values in categorical variables in high dimensional datasets. The proposed method explicitly uses the information on the association among attributes. Using different simulation settings, the performance is compared with available imputation methods. A variety of real data sets, containing heart, DNA, and lymphatic cancer, is also used to support the results obtained by simulations. The results show that the weighting of attributes yields smaller imputation errors than existing approaches like random forest and MICE.

Adapting Nearest Neighbor for Multiple Imputation: Advantages, Challenges, and Drawbacks

Abstract The U.S. Census Bureau has historically used nearest neighbor (NN) or random hot deck (RHD) imputation to handle missing data for many types of survey data. Using these methods removes the need to parametrically model values in imputation models. With strong auxiliary information, NN imputation is preferred because it produces more precise estimates than RHD. In addition, NN imputation is robust against a misspecified response mechanism if missingness depends on the auxiliary variable, in contrast to RHD which ignores the auxiliary information. A compromise between these two methods is k-NN imputation, which identifies a set of the k closest neighbors (“donor pool”) and randomly selects a single donor from this set. Recently these methods have been used for multiple imputation (MI), enabling variance estimation via the so-called Rubin’s Combining Rules. The Approximate Bayesian Bootstrap (ABB) is a simple-to-implement algorithm that makes the RHD “proper” for MI. In concept, ABB should work to propagate uncertainty for NN MI; bootstrapping respondents mean each nonrespondent’s one “nearest” donor will not be available for every imputation. However, we demonstrate through simulation that NN MI using ABB leads to variance underestimation. This underestimation is somewhat but not entirely attenuated with k-NN imputation. An alternative approach to variance estimation after MI, bootstrapped MI, eliminates the underestimation with NN imputation, but we show that it suffers from overestimation of variance with nonnegligible sampling fractions under both equal and unequal probability sampling designs. We propose a modification to bootstrapped MI to account for nonnegligible sampling fractions. We compare the performance of RHD and the various NN MI methods under a variety of sampling designs, sampling fractions, distribution shapes, and missingness mechanisms.

A Bayesian network approach incorporating imputation of missing data enables exploratory analysis of complex causal biological relationships.

Bayesian networks can be used to identify possible causal relationships between variables based on their conditional dependencies and independencies, which can be particularly useful in complex biological scenarios with many measured variables. Here we propose two improvements to an existing method for Bayesian network analysis, designed to increase the power to detect potential causal relationships between variables (including potentially a mixture of both discrete and continuous variables). Our first improvement relates to the treatment of missing data. When there is missing data, the standard approach is to remove every individual with any missing data before performing analysis. This can be wasteful and undesirable when there are many individuals with missing data, perhaps with only one or a few variables missing. This motivates the use of imputation. We present a new imputation method that uses a version of nearest neighbour imputation, whereby missing data from one individual is replaced with data from another individual, their nearest neighbour. For each individual with missing data, the subsets of variables to be used to select the nearest neighbour are chosen by sampling without replacement the complete data and estimating a best fit Bayesian network. We show that this approach leads to marked improvements in the recall and precision of directed edges in the final network identified, and we illustrate the approach through application to data from a recent study investigating the causal relationship between methylation and gene expression in early inflammatory arthritis patients. We also describe a second improvement in the form of a pseudo-Bayesian approach for upweighting certain network edges, which can be useful when there is prior evidence concerning their directions.

Imputation methods for high-dimensional mixed-type datasets by nearest neighbors

In modern biomedical research, the data often contain a large number of variables of mixed data types (continuous, multi-categorical, or binary) but on some variables observations are missing. Imputation is a common solution when the downstream analyses require a complete data matrix. Several imputation methods are available that work under specific distributional assumptions. We propose an improvement over the popular non-parametric nearest neighbor imputation method which requires no particular assumptions. The proposed method makes practical and effective use of the information on the association among the variables. In particular, we propose a weighted version of the Lq distance for mixed-type data, which uses the information from a subset of important variables only. The performance of the proposed method is investigated using a variety of simulated and real data from different areas of application. The results show that the proposed methods yield smaller imputation error and better performance when compared to other approaches. It is also shown that the proposed imputation method works efficiently even when the number of samples is smaller than the number of variables.

An improved pollution forecasting model with meteorological impact using multiple imputation and fine-tuning approach

Air pollution forecasting is a significant step for air quality pollution management to mitigate pollution’s negative impact on the environment and people’s health. The data-driven forecasting model can help a better understanding of environmental air quality. The existing data-driven forecasting models usually ignore missing values, the correlations between the pollutant and meteorological factors and fail to perform temporal modeling effectively, affecting prediction accuracy. In response to these issues, we present a deep learning-based Convolutional LSTM–SDAE (CLS) model to forecast the particulate matter level, revealing the correlation between particulate matter and meteorological factors. In the proposed architecture, the k nearest neighbor (KNN) imputation technique is employed to recover the air quality dataset’s missing values. The Convolutional Long Short Term Memory (CNN–LSTM) unit identifies the vast dataset’s hidden features and performs pollutants’ temporal modeling. In addition, Bidirectional Gatted Recurrent Unit (BIGRU) is implemented as both encoder and decoder in Sparse Denoising Autoencoder, which reconstructs the CNN–LSTM model’s output in the dynamic fine-tuning layer to get robust prediction results. The experimental results in Talcher, India, and Beijing, China indicate that the model can improve forecasting accuracy and outperforms the other state of art and baseline models.

R-HEFS: Rough set based heterogeneous ensemble feature selection method for medical data classification

Feature selection is one of the trustworthy processes of dimensionality reduction technique to select a subset of relevant and non-redundant features from large datasets. Ensemble feature selection (EFS) approach is a recent technique aiming at accumulating diversity in the subset of selected features. It improves the performance of learning algorithms and obtains more stable and robust results. In this paper, a novel rough set theory (RST) based heterogeneous EFS method (R-HEFS) is proposed for selecting the less redundant and highly relevant features during the aggregation of diverse feature subsets by applying the feature-class, feature-feature rough dependency and feature-significance measures. In R-HEFS five state-of-the-art RST based filter methods are used as a base feature selectors. Experiments are carried out on 10 benchmark medical datasets collected from the UCI repository. For the imputation of the missing values and discretization of the continuous features, k nearest neighbor (kNN) imputation method and RST based discretization techniques are applied. The effectiveness of the proposed R-HEFS method is evaluated and analyzed by using four benchmark classifiers viz., Naïve Bayes (NB), random forest (RF), support vector machine (SVM), and AdaBoost. The proposed R-HEFS method turns out to be effective by removing the non-relevant and redundant features during the process of aggregation of base feature selectors and it assists to increase the classification accuracy. Out of 10 different medical datasets, on 7 datasets, R-HEFS has achieved better average classification accuracy. So, the overall results strongly suggest that the proposed R-HEFS method can reduce the dimension of large medical datasets and may help the physicians or medical experts to diagnose (classify) different diseases with lesser computational complexities.