Scikit-learn Library Research Articles

Recyclable magnetic orange peel residues modified by anionic surfactant for basic blue 9 removal: Experimental study and machine learning modeling

Three successive treatments were applied to modify the orange peel (OP) residues using NaOH, sodium dodecyl sulfate (SDS), and Fe3O4 nanoparticles for the removal of basic blue 9 (BB9). The structure and morphology of the chemically modified adsorbents were investigated using various analytical techniques including PXRD, FTIR, FESEM/EDAX, VSM, and N2 physisorption analyses. The removal of BB9 by modified adsorbents was modeled using two machine learning (mL) approaches. A gradient boosting regression tree (GBRT) model for 8 adsorbents and extreme gradient boosting (XGBoost) model for optimal adsorbent were designed based on 584 and 73 sets of experimental data, respectively. The scikit-learn library was used to determine the importance of each feature for building the GBRT and XGBoost models and it was confirmed that all the parameters affect the BB9 removal process to a remarkable extend. The adsorption isotherm data were fitted to Langmuir, Freundlich, and Temkin models. The kinetics and thermodynamics investigations showed that the removal of BB9 follows a pseudo-second-order kinetic model that is spontaneous and exothermic. The BB9 desorption experiments in water and HCl showed that the electrostatic interaction has a significant role in the removal of BB9 on chemically modified orange peel residues. The adsorption mechanism studies using Boyd’s theory indicated that the adsorption of BB9 on the modified adsorbents was mainly governed by the external mass transport where the particle diffusion was the rate limiting step.

Graph neural network modelling as a potentially effective method for predicting and analyzing procedures based on patients' diagnoses

BackgroundCurrently, the healthcare sector strives to improve the quality of patient care management and to enhance/increase its economic performance/efficiency (e.g., cost-effectiveness) by healthcare providers. The data stored in electronic health records (EHRs) offer the potential to uncover relevant patterns relating to diseases and therapies, which in turn could help identify empirical medical guidelines to reflect best practices in a healthcare system. Based on this pattern of identification model, it is thus possible to implement recommender systems with the notion that a higher volume of procedures is often associated with better high-quality models. MethodsAlthough there are several different applications that uses machine learning methods to identify such patterns, such identification is still a challenge, due in part because these methods often ignore the basic structure of the population, or even considering the similarity of diagnoses and patient typology. To this end, we have developed a method based on graph-data representation aimed to cluster ‘similar’ patients. Using such a model, patients will be linked when there is a same and/or similar patterns are being observed amongst them, a concept that will enable the construction of a network-like structure which is called a patient graph.11In several fields in science the concept of network is preferred over the concept of graph, which is commonly employed in informatics. In this article we will continue using the concept of graph due to the fact that the implemented methodology relies on methods coming from informatics but we will point out to the concept of network when we explicitly require a more visual definition that considers nodes interlinked with edges. This structure can be then analyzed by Graph Neural Networks (GNN) to identify relevant labels, and in this case the appropriate medical procedures that will be recommended. ResultsWe were able to construct a patient graph structure based on the patient's basic information like age and gender as well as the diagnosis and the trained GNNs models to identify the corresponding patient's therapies using a synthetic patient database. We have even compared our GNN models against different baseline models (using the SCIKIT-learn library of python) and also against the performance of these different model-methods. We have found that the GNNs models are superior, with an average improvement of the f1 score of 6.48 % in respect to the baseline models. In addition, the GNNs models are useful in performing additional clustering analysis which allow a distinctive identification of specific therapeutic/treatment clusters relating to a particular combination of diagnoses. ConclusionsWe found that the GNNs models offer a promising lead to model the distribution of diagnoses in patient population, and is thus a better model in identifying patients with similar phenotype based on the combination of morbidities and/or comorbidities. Nevertheless, network/graph building is still challenging and prone to biases as it is highly dependent on how the ICD distribution affects the patient network embedding space. This graph setup not only requires a high quality of the underlying diagnostic ecosystem, but it also requires a good understanding on how patients at hand are identified by disease respectively. For this reason, additional work is still needed to better improve patient embedding in graph structures for future investigations and the applications of this service-based technology. Therefore, there has not been any interventional study yet.

Understanding and Predicting Cognitive Improvement of Young Adults in Ischemic Stroke Rehabilitation Therapy.

Accurate early predictions of a patient's likely cognitive improvement as a result of a stroke rehabilitation programme can assist clinicians in assembling more effective therapeutic programs. In addition, sufficient levels of explainability, which can justify these predictions, are a crucial requirement, as reported by clinicians. This article presents a machine learning (ML) prediction model targeting cognitive improvement after therapy for stroke surviving patients. The prediction model relies on electronic health records from 201 ischemic stroke surviving patients containing demographic information, cognitive assessments at admission from 24 different standardized neuropsychology tests (e.g., TMT, WAIS-III, Stroop, RAVLT, etc.), and therapy information collected during rehabilitation (72,002 entries collected between March 2007 and September 2019). The study population covered young-adult patients with a mean age of 49.51 years and only 4.47% above 65 years of age at the stroke event (no age filter applied). Twenty different classification algorithms (from Python's Scikit-learn library) are trained and evaluated, varying their hyper-parameters and the number of features received as input. Best-performing models reported Recall scores around 0.7 and F1 scores of 0.6, showing the model's ability to identify patients with poor cognitive improvement. The study includes a detailed feature importance report that helps interpret the model's inner decision workings and exposes the most influential factors in the cognitive improvement prediction. The study showed that certain therapy variables (e.g., the proportion of memory and orientation executed tasks) had an important influence on the final prediction of the cognitive improvement of patients at individual and population levels. This type of evidence can serve clinicians in adjusting the therapeutic settings (e.g., type and load of therapy activities) and selecting the one that maximizes cognitive improvement.

IMPLEMENTASI METODE DATA MINING K-MEANS CLUSTERING TERHADAP DATA PEMBAYARAN TRANSAKSI MENGGUNAKAN BAHASA PEMROGRAMAN PYTHON PADA CV DIGITAL DIMENSI

CV Digital Dimensi is a company engaged in the printing sector, which is a subsidiary of the XG Group located in Jakarta. In order to be able to compete with other companies, the company does not only focus on products and services but also efforts to build personal relationships with customers. Existing payment transaction data so far have not been utilized as much as possible in determining the company's business strategy, therefore the implementation of data mining is necessary to analyze and explore the available data to find new information that is more valuable and useful for the company. The K-Means Clustering method is a data mining technique to obtain data groups by maximizing the similarity of characteristics within the cluster and maximizing the differences between clusters. The purpose of this study is to apply the K-Means Clustering method to group sales transaction data on CV Digital Dimensi and display the results in the form of visual graphics using the Python programming language and Scikit-Learn library. The results of this study succeeded in classifying sales transaction data into five clusters and can be used as a reference in determining the company's business strategy.

Fake News Detection Using Machine Learning

Abstract: The role of social media in our day to day life has increased rapidly in recent years. Information quality in social media is an increasingly important issue, but web-scale data hinders experts’ ability to assess and correct much of the inaccurate content, or “fake news”, present in these platforms. It is now used not only for social interaction, but also as an important platform for exchanging information and news. Twitter, Facebook a micro-blogging service, connects millions of users around the world and allows for the real-time propagation of information and news. The fake news on social media and various other media is wide spreading and is a matter of serious concern due to its ability to cause a lot of social and national damage with destruction impacts. A lot of research is already focused on detecting it. A human being is unable to detect all these fake news. Detecting fake news is an important step. This process will result in feature extraction and vectorization; we propose using Python scikit-learn library to perform tokenization and feature extraction of text data, because this library contains useful tools like Count Vectorizer and Tiff Vectorizer. Then, we will perform feature selection methods, to experiment and choose the best fit features to obtain the highest precision, according to confusion matrix results. A feature analysis then identifies features that are most predictive for crowdsourced and journalistic accuracy assessments, results of which are consistent with prior work. We aim to provide the user with the ability to classify the news as “fake” or “real”.

A new multi-label dataset for Web attacks CAPEC classification using machine learning techniques

ContextThere are many datasets for training and evaluating models to detect web attacks, labeling each request as normal or attack. Web attack protection tools must provide additional information on the type of attack detected, in a clear and simple way. ObjectivesThis paper presents a new multi-label dataset for classifying web attacks based on CAPEC classification, a new way of features extraction based on ASCII values, and the evaluation of several combinations of models and algorithms. MethodsUsing a new way to extract features by computing the average of the sum of the ASCII values of each of the characters in each field that compose a web request, several combinations of algorithms (LightGBM and CatBoost) and multi-label classification models are evaluated, to provide a complete CAPEC classification of the web attacks that a system is suffering. The training and test data used for training and evaluating the models come from the new SR-BH 2020 multi-label dataset. ResultsCalculating the average of the sum of the ASCII values of the different characters that make up a web request shows its usefulness for numeric encoding and feature extraction. The new SR-BH 2020 multi-label dataset allows the training and evaluation of multi-label classification models, also allowing the CAPEC classification of the various attacks that a web system is undergoing. The combination of the two-phase model with the MultiOutputClassifier module of the scikit-learn library, together with the CatBoost algorithm shows its superiority in classifying attacks in the different criticality scenarios. ConclusionExperimental results indicate that the combination of machine learning algorithms and multi-phase models leads to improved prediction of web attacks. Also, the use of a multi-label dataset is suitable for training learning models that provide information about the type of attack.

Estimation of gas turbine technical condition using machine learning methods

This paper considers the method to estimate the technical condition of gas turbine power for natural gas transportation, using machine learning methods. Source data was used to archive gas-dynamic parameters from the automatic control system of the gas turbine. The method is based on changing the enthalpy of the natural gas before and after the centrifugal gas compressor is used for creating a dataset with measured parameters and power from the gas turbine. The actual power is determined from the line of modes for a certain period. The software is implemented using Python and the Scikit-learn library is used to create machine learning models. A mean average percentile error is chosen as the model quality criterion. In this paper, different sets of feature parameters and sample sizes are researched by the quality of the prediction machine learning models. Recommendations on the use of models are given. It has been established that the approach is not applicable for predicting future technical condition without the presence of data on a similar technical condition in the training sample. It is recommended to use the described approach to determine the technical condition in a period of operation in the past.

Breast Cancer Prediction using KNN, SVM, Logistic Regression and Decision Tree

Abstract: Each year number of deaths is increasing extremely because of breast cancer. It is the most frequent type of all cancers and the major cause of death in women worldwide. Any development for prediction and diagnosis of cancer disease is capital important for a healthy life. Consequently, high accuracy in cancer prediction is important to update the treatment aspect and the survivability standard of patients. Machine learning techniques can bring a large contribute on the process of prediction and early diagnosis of breast cancer, became a research hotspot and has been proved as astrong technique. In this study, we applied five machine learning algorithms: Support Vector Machine (SVM), RandomForest, Logistic Regression, Decision tree (C4.5) and K-Nearest Neighbours (KNN) on the Breast Cancer WisconsinDiagnostic dataset, after obtaining the results, a performance evaluation and comparison is carried out between these different classifiers. The main objective of thisresearch paperisto predict and diagnosis breast cancer, using machine- learning algorithms, and find out the most effective whit respect to confusion matrix, accuracy and precision. It is observed that Support vector Machine outperformed all other classifiers and achieved the highest accuracy (97.2%). All the work is done in the Anaconda environment based on python programming language and Scikit-learn library. Keywords: Decision Tree, KNN, SVM, Malignant, Benign , Logistic Regression

Fake News Detection Using Machine Learning

ABSTRACT–False stories on social media and various other platforms spread rumors and are a source of great concern as they have the potential to wreak havoc on the social and social ills with devastating consequences. A lot of research is going on just to focus on finding it. This paper analyzes research related to the detection of false information and can examine the culture of machine learning models to select the best to create a product model with a supervised machine learning algorithm, which can distinguish or find false stories true or false, using tools such as python scikit-learn, i -NPL text analysis. This process will lead to feature removal and vectorization; we should use the python scikit-learn library to make tokens and extract the text data feature, as this library contains useful tools such as Count Vectorizer and Tiff Vectorizer. Then we will create methods to select test features and select advanced file features to get the highest accuracy, depending on the results of the confusion matrix. Keyword: count vectorizer, NLP, precision.

Performance Analysis of different Machine Learning Models for Intrusion Detection Systems

In recent years, the world witnessed a rapid growth in attacks on the internet which resulted in deficiencies in networks performances. The growth was in both quantity and versatility of the attacks. To cope with this, new detection techniques are required especially the ones that use Artificial Intelligence techniques such as machine learning based intrusion detection and prevention systems. Many machine learning models are used to deal with intrusion detection and each has its own pros and cons and this is where this paper falls in, performance analysis of different Machine Learning Models for Intrusion Detection Systems based on supervised machine learning algorithms. Using Python Scikit-Learn library KNN, Support Vector Machine, Naïve Bayes, Decision Tree, Random Forest, Stochastic Gradient Descent, Gradient Boosting and Ada Boosting classifiers were designed. Performance-wise analysis using Confusion Matrix metric carried out and comparisons between the classifiers were a due. As a case study Information Gain, Pearson and F-test feature selection techniques were used and the obtained results compared to models that use all the features. One unique outcome is that the Random Forest classifier achieves the best performance with an accuracy of 99.96% and an error margin of 0.038%, which supersedes other classifiers. Using 80% reduction in features and parameters extraction from the packet header rather than the workload, a big performance advantage is achieved, especially in online environments.

Machining performance analysis for PMEDM of biocompatible material Ti-6Al-7Nb alloy: A machine learning approach

Surface properties of biomedical implants have vital role as they release toxic particles due to corrosion and wear under repeated loading. Powder mixed electric discharge machining (PMEDM) is used to shape difficult-to-machine materials with simultaneous surface modification. In the present work Taguchi’s L27 orthogonal array is used to machine Ti-6Al-7Nb titanium alloy using PMEDM. SiC powder of size < 53 µm is added in varying concentrations to the dielectric. Input parameters chosen for the investigation are powder concentration (Cp), peak current (Ip), pulse on time (Ton), duty cycle (DC) and gap voltage (Vg). The influence of input parameters on material removal rate (MRR), surface roughness (Ra) and tool wear rate (TWR) has been investigated. Machine learning algorithm has been used for data analysis, processing and a prediction model has been established by applying linear regression using python programming language. Numpy, Pandas libraries of python language has been used for data analysis, data pre-processing and prediction has been done using Scikit-Learn library. The error between predicted and confirmatory test results is 2.14%.

Common Laboratory Parameters Are Useful for Screening for Alcohol Use Disorder: Designing a Predictive Model Using Machine Learning.

The diagnosis of alcohol use disorder (AUD) remains a difficult challenge, and some patients may not be adequately diagnosed. This study aims to identify an optimum combination of laboratory markers to detect alcohol consumption, using data science. An analytical observational study was conducted with 337 subjects (253 men and 83 women, with a mean age of 44 years (10.61 Standard Deviation (SD)). The first group included 204 participants being treated in the Addictive Behaviors Unit (ABU) from Albacete (Spain). They met the diagnostic criteria for AUD specified in the Diagnostic and Statistical Manual of mental disorders fifth edition (DSM-5). The second group included 133 blood donors (people with no risk of AUD), recruited by cross-section. All participants were also divided in two groups according to the WHO classification for risk of alcohol consumption in Spain, that is, males drinking more than 28 standard drink units (SDUs) or women drinking more than 17 SDUs. Medical history and laboratory markers were selected from our hospital’s database. A correlation between alterations in laboratory markers and the amount of alcohol consumed was established. We then created three predicted models (with logistic regression, classification tree, and Bayesian network) to detect risk of alcohol consumption by using laboratory markers as predictive features. For the execution of the selection of variables and the creation and validation of predictive models, two tools were used: the scikit-learn library for Python, and the Weka application. The logistic regression model provided a maximum AUD prediction accuracy of 85.07%. Secondly, the classification tree provided a lower accuracy of 79.4%, but easier interpretation. Finally, the Naive Bayes network had an accuracy of 87.46%. The combination of several common biochemical markers and the use of data science can enhance detection of AUD, helping to prevent future medical complications derived from AUD.

RPkNN: An OpenCL-Based FPGA Implementation of the Dimensionality-Reduced kNN Algorithm Using Random Projection

Due to the so-called curse of dimensionality and increase in the size of databases, there is an ever-increasing demand for computing resources and memory bandwidth when performing the k-nearest neighbors (kNNs) algorithm, resulting in a slow-down to process large datasets. This work presents an OpenCL-based framework for accelerating the kNN algorithm on field-programmable gate arrays (FPGAs) benefiting from the random projection dimensionality reduction. The proposed RPkNN framework includes two compute modules implementing a throughput-optimized hardware architecture based on random projection and the kNN algorithm and a host program facilitating easy integration of the compute modules in the existing applications. RPkNN also utilizes a new buffering scheme tailored to random projection and the kNN algorithm. The proposed architecture enables parallel kNN computations with a single memory channel and takes advantage of the sparsity features of the input data to implement a highly optimized and parallel implementation of random projection. We employ a computation storage device (CSD) to directly access the high-dimensional data on non-volatile memory express (NVMe) solid state drive (SSD) and store and reuse the compressed and low-dimensional data on the FPGA dynamic random access memory (DRAM), hence eliminating data transfers to the host DRAM. We compare RPkNN implemented on the Samsung SmartSSD CSD with the kNN implementation of the scikit-learn library running on an Intel Xeon Gold 6154 CPU. The experimental results show that the proposed RPkNN solution achieves, on average, <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$26\times $ </tex-math></inline-formula> and <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$46\times $ </tex-math></inline-formula> higher performance across different dimensions per a single kNN computation for the SIFT1M and GIST1M databases, respectively. Finally, RPkNN is <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$1.7\times $ </tex-math></inline-formula> faster than the similar FPGA-based reference method.

33 Setting the Standard: Using a National Burn Registry to Benchmark Risk Adjusted Mortality

IntroductionReports of single center experience and studies of larger databases have identified several predictors of burn center mortality, including age, burn size, and inhalation injury. None of these analyses has been broad enough to allow benchmarking across burn centers. The purpose of this study was to derive a reliable risk-adjusted, statistical model of mortality based on real-life experience at many of the burn centers in the United States.MethodsWe used a national burn registry to identify 128,252 initial admissions from July 2015 through June 2020 across 103 unique burn centers. We selected 23 predictor variables, from over 50 recorded in the dataset based on completeness (at least 75% complete required) and clinical significance. Missing data were multiply imputed with a Bayesian Ridge Regression estimator. All data analysis was performed in Python using Numpy and Scikit-Learn libraries. We used Gradient boosted regression (CatBoost), a form of machine learning, to predict mortality and compared this to traditional logistic regression. Model performance was evaluated with AUC and PR curves. Using the CatBoost predictions, observe to expected mortality was calculated for each center. Confidence intervals for O/E analysis in the case of mortality prediction were calculated using a custom implementation of the Clopper-Pearson method. Analyses were run on three cohorts: All patients; Patients with 10-20% TBSA; and >20% TBSA.ResultsThe CatBoost model achieved a test AUC of 0.982 with an average precision of 0.801. The logistic regression, by comparison, produced an AUC of 0.974 with an average precision of 0.726. While accuracy, the measure most reported in the literature, is near ceiling for both models, the CatBoost model is markedly more sensitive, leading to a substantial improvement in average precision. Because of the superiority of the CatBoost model with respect to outcome prediction, we only used CatBoost models for calculation of O/E ratio (Fig. 1).ConclusionsGradient boosted regression models provided greater model performance than traditional, multivariate, logistic regression. Using data from a national burn registry, we can predict burn mortality across contributing centers allowing for meaningful O/E ratios. Further, this allows for comparison of mortality across centers contributing data to the registry.

34 Using a National Burn Registry to Develop a Model for Risk-adjusted Length of Stay Benchmarking

IntroductionLength of stay (LOS) is a frequently reported outcome after a burn injury. Previous literature estimates LOS at 1 day per % total burn surface area (TBSA) but this varies considerably across patients & centers. LOS benchmarking will benefit individual burn centers as a way to measure their performance & set expectations for patients. We sought to create a nationwide, risk-adjusted model to allow for LOS benchmarking based on data from a national burn registry.MethodsUsing data from a national burn registry, we queried admissions from 7/2015-6/2020 & identified 126,129 records with LOS data reported by 103 centers. We selected 23 predictor variables on the basis of completeness (min. 75% required) & clinical significance. Missing data were multiply imputed with a Bayesian Ridge Regression estimator. All statistics were calculated in Python using Numpy & Scikit-Learn libraries. Comparisons of unpenalized linear regression & Gradient boosted (CatBoost) regressor models were performed by measuring the R2 & concordance correlation coefficient (CCC) on the application of the model to the test dataset. The CatBoost model applied to bootstrapped versions of the entire dataset was then used to calculate O/E ratios for individual burn centers. Confidence intervals (CI) for O/E ratios were calculated using a normal distribution parametric model. Analyses were run on 3 cohorts: all patients, 10-20% TBSA, >20% TBSA.ResultsThe CatBoost model outperformed the linear regression model with a test R2 of 0.68 & CCC of 0.81 compared to the regression model with R2=0.52, CCC=0.70. The CatBoost was also less biased for higher & lower LOS durations. Due to the CatBoost model’s superiority in predicting the outcome, this model alone was used for O/E ratio calculations. The O/E ratio data from the model for all 3 cohorts are shown in Figure 1.ConclusionsGradient boosted regression models provided greater model performance than traditional regression analysis. Using national burn data, we can predict LOS across contributing burn centers while accounting for patient & center characteristics, producing more meaningful O/E ratios. These models provide a risk-adjusted LOS benchmarking using a robust data source, the first of its kind, for burn centers.

Hybrid Missing Value Imputation Algorithm- KLR

The data mining process mainly deals with the estimation, prediction, pattern extraction, and classification in big databases. The presence of missing values in the dataset decreases the accuracy of data mining classifiers. Therefore it is necessary to deal with missing values in the dataset to achieve accurate results. To improve the quality of data and prediction accuracy in the classification process, the authors have proposed a new hybrid missing value prediction algorithm, KLR, by combining the KNN and linear regression approach. The proposed KLR algorithm has been used for class validation and missing values imputation. Wisconsin Breast Cancer Diagnostic Dataset of 569 instances with 32 attributes from the machine learning repository of UCI, Irvinewasused to conduct the study. The Pearson Coefficient Correlation method is used for feature selection.Data normalization is performed using Min-max scaling technique. The Scikit-learn library for machine learning in python is used to complete all the experiments as the experimental framework. The mean square error method is used to evaluate the performance of the model. The proposed KLR algorithm with 450 nearest neighbors out of 569 gives the lowest MSE ie 0.00188 and more accurately predicts the missing values as compared to the classic models.

A176 UTILITY OF MACHINE LEARNING FOR SERUM METABOLOMIC DATA ANALYSIS IN PEDIATRIC CROHN DISEASE

BackgroundThe pathogenesis of pCD remains poorly understood, but evidence suggests roles for genetics, environment, immune response, and gut microbes. Microbial changes can contribute to chronic inflammation and correlate with disease severity. Metabolomics reflects interactions between host immune and gut microbial function by quantifying compounds in biological samples. Therefore, metabolomics provides a unique opportunity to gain insight into pCD pathogenesis.AimsTo correlate disease severity, metabolites, and clinical data by applying machine learning algorithms in pediatric Crohn Disease (pCD).MethodsImageKids is a multicenter, prospective, cohort observational study, conducted to develop magnetic resonance enterography (MRE) indices for pCD. Paired serum specimens were collected at study initiation (Visit One; V1) and completion (Visit Four; V4; 18 months) for 120 pCD patients. Serum from patients with representative clinical scenarios and paired samples was analyzed at The Metabolomics Innovation Centre (TMIC; University of Alberta) and 131 metabolites were identified. Metabolites were analyzed via Unsupervised (U.ML) and Supervised (S.ML) Machine Learning algorithms based on Scikit-learn library in Python. Principal Component Analysis (PCA) was used to identify the variation pattern of the patients’ metabolome. Classifiers and regression algorithms were trained to assess correlation with disease activity.ResultsResults were available for the 56 paired samples. U.ML demonstrated distinct metabolome profiles with V1 clustering mainly attributed to aspartic acid, glutamic acid, and kynurenine. V4 clustering was mainly attributed to spermidine, spermine, total dimethylarginine. Furthermore, demographics was found as an important environmental factor driving distinct patterns of the metabolomics profile.After training different classifiers and regressors with S. ML algorithms, metabolome data were correlated with disease severity (defined by C-reactive protein and fecal calprotectin). Isoleucine, p-hydroxyhippuric acid, and putrescine were the top three compounds associated with disease severity. The accuracy of our classification models was of 80% and the coefficient of determination of our regression models was 0.5ConclusionsMetabolomic analysis can provide insight into disease pathogenesis and help predict disease severity among pCD patients. The correlation between metabolomics and disease severity might allow a better understanding of changes in host-microbe interactions and introduce new diagnostic or therapeutic options.Funding AgenciesCIHR

Investigation of the Effect of Normalization Methods on ANFIS Success: Forestfire and Diabets Datasets

Machine learning and artificial intelligence techniques are more and more in our lives and studies in this field are increasing day by day. Data is vital for these studies. In order to draw meaningful conclusions from the available data, new methods are proposed and successful results are obtained. The preparation of the obtained data is very important in the studies to be carried out. Data preprocessing is very important in the preparation of data. The most critical stage of the data preprocessing process is the scaling or normalization of the data. Machine learning libraries such as scikit-learn and programming languages such as R provide the necessary libraries to scale data. However, it is not known exactly which normalization method will be applied and which will yield more successful results. The success of these normalization methods has been investigated on many different methods, but such a study has not been done on the adaptive neural fuzzy inference system (ANFIS). The aim of this study is to examine the success of normalization methods on ANFIS in terms of both classification and regression problems. So, for studies using the Anfis method, guidance will be provided on which normalization process will give better results in the data preprocessing stage. Four different normalization methods in the scikit-learn library were applied on the Diabets and Forestfire datasets in the UCI database. The results are presented separately for both classification and regression. It has been determined that min-max normalization in classification problems and working with original data in regression problems are more successful.

Predicting the subclinical carotid atherosclerosis in overweight and obese patients using a machine learning model

Aim. To develop a model for predicting the subclinical carotid atherosclerosis (SCA) in order to refine cardiovascular risk (CVR) using machine learning methods in overweight and obese patients without hypertension, diabetes and/or cardiovascular disease (CVD).Material and methods. Anonymized database (DB) Webiomed (2.9 million patients) was used. There were following inclusion criteria: age ≥18 years, body mass index ≥25 kg/m2, availability of data on ultrasound of extracranial arteries. Patients with hypertension, diabetes and/or CVD were excluded from the analysis. Data on 5750 patients were selected, of which atherosclerotic plaques were detected in 385 people. The final data set contained information on 447 patients, 197 (44,1%) of which had SCA. Quantitative and categorical traits for model training were taken with 40% occupancy in the database. The number of final traits for machine learning was 28. When creating the model, 3 Random Forest algorithms, AdaBoostClassifier, KNeighborsClassifier and the Scikit-learn library were used. To improve the model performance, the fill missing function was used. The target parameters of the model were given a predictive ability (accuracy) of at least 75%, while the area under the ROC curve was at least 0,75.Results. The resulting dataset was divided into training and test parts in a ratio of 80:20. Depending on the applied algorithms, the learned model was characterized by a predictive ability of 75-97%, sensitivity of 77-92%, specificity of 80-98%, and area under the ROC-curve of 0,88-0,97. Taking into account the accuracy metrics, the best results were obtained for the model learned by the Random Forest algorithm (95%, 92%, 98% and 0,95, respectively).Conclusion. The developed model can help a physician make a decision to refer an overweight and obese patient without cardiovascular diseases for ultrasound of extracranial arteries, which contributes to a more accurate CVR stratification. The introduction of such risk stratification algorithms into practice will increase the accuracy and quality of CVR prediction and optimize the system of preventive measures.

Predicting the subclinical carotid atherosclerosis in overweight and obese patients using a machine learning model

Aim. To develop a model for predicting the subclinical carotid atherosclerosis (SCA) in order to refine cardiovascular risk (CVR) using machine learning methods in overweight and obese patients without hypertension, diabetes and/or cardiovascular disease (CVD).Material and methods. Anonymized database (DB) Webiomed (2.9 million patients) was used. There were following inclusion criteria: age ≥18 years, body mass index ≥25 kg/m2, availability of data on ultrasound of extracranial arteries. Patients with hypertension, diabetes and/or CVD were excluded from the analysis. Data on 5750 patients were selected, of which atherosclerotic plaques were detected in 385 people. The final data set contained information on 447 patients, 197 (44,1%) of which had SCA. Quantitative and categorical traits for model training were taken with 40% occupancy in the database. The number of final traits for machine learning was 28. When creating the model, 3 Random Forest algorithms, AdaBoostClassifier, KNeighborsClassifier and the Scikit-learn library were used. To improve the model performance, the fill missing function was used. The target parameters of the model were given a predictive ability (accuracy) of at least 75%, while the area under the ROC curve was at least 0,75.Results. The resulting dataset was divided into training and test parts in a ratio of 80:20. Depending on the applied algorithms, the learned model was characterized by a predictive ability of 75-97%, sensitivity of 77-92%, specificity of 80-98%, and area under the ROC-curve of 0,88-0,97. Taking into account the accuracy metrics, the best results were obtained for the model learned by the Random Forest algorithm (95%, 92%, 98% and 0,95, respectively).Conclusion. The developed model can help a physician make a decision to refer an overweight and obese patient without cardiovascular diseases for ultrasound of extracranial arteries, which contributes to a more accurate CVR stratification. The introduction of such risk stratification algorithms into practice will increase the accuracy and quality of CVR prediction and optimize the system of preventive measures.