scholarly journals Predicting Kidney Graft Survival using Machine Learning Methods: Prediction Model Development and Feature Significance Analysis Study (Preprint)

10.2196/26843 ◽  
2021 ◽  
Author(s):  
Syed Asil Ali Naqvi ◽  
Karthik Tennankore ◽  
Amanda Vinson ◽  
Patrice C. Roy ◽  
Syed Sibte Raza Abidi
Keyword(s):  
Machine Learning ◽  
Prediction Model ◽  
Graft Survival ◽  
Model Development ◽  
Kidney Graft ◽  
Learning Methods ◽  
Analysis Study ◽  
Machine Learning Methods ◽  
Significance Analysis ◽  
Feature Significance

scholarly journals Comparison of Soil Total Nitrogen Content Prediction Models Based on Vis-NIR Spectroscopy

Sensors ◽  
2020 ◽  
Vol 20 (24) ◽  
pp. 7078
Author(s):  
Yueting Wang ◽  
Minzan Li ◽  
Ronghua Ji ◽  
Minjuan Wang ◽  
Lihua Zheng
Keyword(s):  
Machine Learning ◽  
Deep Learning ◽  
Prediction Model ◽  
Total Nitrogen ◽  
Nir Spectroscopy ◽  
Learning Method ◽  
Learning Methods ◽  
Data Set ◽  
Machine Learning Methods ◽  
Conventional Machine

Visible-near-infrared spectrum (Vis-NIR) spectroscopy technology is one of the most important methods for non-destructive and rapid detection of soil total nitrogen (STN) content. In order to find a practical way to build STN content prediction model, three conventional machine learning methods and one deep learning approach are investigated and their predictive performances are compared and analyzed by using a public dataset called LUCAS Soil (19,019 samples). The three conventional machine learning methods include ordinary least square estimation (OLSE), random forest (RF), and extreme learning machine (ELM), while for the deep learning method, three different structures of convolutional neural network (CNN) incorporated Inception module are constructed and investigated. In order to clarify effectiveness of different pre-treatments on predicting STN content, the three conventional machine learning methods are combined with four pre-processing approaches (including baseline correction, smoothing, dimensional reduction, and feature selection) are investigated, compared, and analyzed. The results indicate that the baseline-corrected and smoothed ELM model reaches practical precision (coefficient of determination (R2) = 0.89, root mean square error of prediction (RMSEP) = 1.60 g/kg, and residual prediction deviation (RPD) = 2.34). While among three different structured CNN models, the one with more 1 × 1 convolutions preforms better (R2 = 0.93; RMSEP = 0.95 g/kg; and RPD = 3.85 in optimal case). In addition, in order to evaluate the influence of data set characteristics on the model, the LUCAS data set was divided into different data subsets according to dataset size, organic carbon (OC) content and countries, and the results show that the deep learning method is more effective and practical than conventional machine learning methods and, on the premise of enough data samples, it can be used to build a robust STN content prediction model with high accuracy for the same type of soil with similar agricultural treatment.

scholarly journals Global challenges of students dropout: A prediction model development using machine learning algorithms on higher education datasets

2021 ◽  
Vol 129 ◽  
pp. 09001
Author(s):  
Meseret Yihun Amare ◽  
Stanislava Simonova
Keyword(s):  
Higher Education ◽  
Machine Learning ◽  
Prediction Model ◽  
Higher Education Institutions ◽  
Retention Rates ◽  
Quality Of Education ◽  
Learning Methods ◽  
Machine Learning Methods ◽  
Global Challenges

Research background: In this era of globalization, data growth in research and educational communities have shown an increase in analysis accuracy, benefits dropout detection, academic status prediction, and trend analysis. However, the analysis accuracy is low when the quality of educational data is incomplete. Moreover, the current approaches on dropout prediction cannot utilize available sources. Purpose of the article: This article aims to develop a prediction model for students’ dropout prediction using machine learning techniques. Methods: The study used machine learning methods to identify early dropouts of students during their study. The performance of different machine learning methods was evaluated using accuracy, precision, support, and f-score methods. The algorithm that best suits the datasets for these performance measurements was used to create the best prediction model. Findings & value added: This study contributes to tackling the current global challenges of student dropouts from their study. The developed prediction model allows higher education institutions to target students who are likely to dropout and intervene timely to improve retention rates and quality of education. It can also help the institutions to plan resources in advance for the coming academic semester and allocate it appropriately. Generally, the learning analytics prediction model would allow higher education institutions to target students who are likely to dropout and intervene timely to improve retention rates and quality of education.

High Performance Concrete (HPC) Compressive Strength Prediction With Advanced Machine Learning Methods

2020 ◽  
pp. 118-140
Author(s):  
Melda Yucel ◽  
Ersin Namlı
Keyword(s):  
Machine Learning ◽  
Compressive Strength ◽  
Prediction Model ◽  
High Performance Concrete ◽  
Prediction Method ◽  
Structural Features ◽  
Strength Prediction ◽  
Concrete Compressive Strength ◽  
Learning Methods ◽  
Machine Learning Methods

In this chapter, prediction applications of concrete compressive strength values were realized via generation of various hybrid models, which are based on decision trees as main prediction method, by using different artificial intelligence and machine learning techniques. In respect to this aim, a literature research was presented. Used machine learning methods were explained together with their developments and structural features. Various applications were performed to predict concrete compressive strength, and then feature selection was applied to prediction model in order to determine primarily important parameters for compressive strength prediction model. Success of both models was evaluated with respect to correct and precision prediction of values with different error metrics and calculations.

Machine learning for CT texture analysis in chemosensitivity of colorectal liver metastases: Initial results.

2019 ◽  
Vol 37 (15_suppl) ◽  
pp. e15148-e15148
Author(s):  
Yuji Miyamoto ◽  
Takeshi Nakaura ◽  
Yukiharu Hiyoshi ◽  
Hideo Baba
Keyword(s):  
Colorectal Cancer ◽  
Machine Learning ◽  
Prediction Model ◽  
Liver Metastases ◽  
Texture Analysis ◽  
Gray Level ◽  
Learning Methods ◽  
Machine Learning Methods ◽  
Chemotherapeutic Response ◽  
Ct Texture Analysis

e15148 Background: Despite advances in cancer treatment over the last decades, more efficacious biomarkers are needed in patients with metastatic colorectal cancer. Several studies have reported that CT texture analysis is a useful prognostic biomarker for patients with colorectal cancer liver metastases (CRLM), however, little study has been done to explore those efficacies using machine learning methods. The present study aimed to evaluate the clinical efficacy of CT texture analysis using machine learning methods as a predictive marker of systemic chemotherapy in patients with CRLM. Methods: Sixty-four patients with CRLM who received first-line chemotherapy were included. Texture analysis was performed on 92 features (First Order Statistics, Gray Level Cooccurrence Matrix, Gray Level Run Length Matrix, Gray Level Size Zone Matrix, Neighbouring Gray Tone Difference Matrix and Gray Level Dependence Matrix) using CT within 1 month before treatment. We evaluated the association between those features and chemotherapeutic response by RECIST (CR+PR vs. SD+PD+NE). We performed eXtreme gradient boost (XGBoost) as a machine learning method to predict the chemotherapeutic response and used the receiver operating characteristic curves to evaluate this prediction model. Results: Main characteristics were the following: male/female = 36/28; median age = 63.5. Patients were treated with oxaliplatin-based chemotherapy (80% of patients), bevacizumab (77%) and anti-EGFR antibody (23%). Thirty-nine patients had confirmed responders, for an overall response rate of 61%, whereas 25 patients (39%) were classified as non-responders (CR: PR: SD: PD: NE = 0: 39: 20: 4: 1). The area under curve of this prediction model was 0.771. Conclusions: We confirmed that CT texture analysis using machine learning for CRLM was feasible. Further analyses are ongoing.

scholarly journals A Zoning Earthquake Casualty Prediction Model Based on Machine Learning

2021 ◽  
Vol 14 (1) ◽  
pp. 30
Author(s):  
Boyi Li ◽  
Adu Gong ◽  
Tingting Zeng ◽  
Wenxuan Bao ◽  
Can Xu ◽  
...  
Keyword(s):  
Machine Learning ◽  
Prediction Model ◽  
Support Vector Regression ◽  
Prediction Method ◽  
Historical Earthquakes ◽  
Support Vector ◽  
Learning Methods ◽  
Support Vector Regression Model ◽  
Machine Learning Methods ◽  
Earthquake Casualty

The evaluation of mortality in earthquake-stricken areas is vital for the emergency response during rescue operations. Hence, an effective and universal approach for accurately predicting the number of casualties due to an earthquake is needed. To obtain a precise casualty prediction method that can be applied to regions with different geographical environments, a spatial division method based on regional differences and a zoning casualty prediction method based on support vector regression (SVR) are proposed in this study. This study comprises three parts: (1) evaluating the importance of influential features on seismic fatality based on random forest to select indicators for the prediction model; (2) dividing the study area into different grades of risk zones with a strata fault line dataset and WorldPop population dataset; and (3) developing a zoning support vector regression model (Z-SVR) with optimal parameters that is suitable for different risk areas. We selected 30 historical earthquakes that occurred in China’s mainland from 1950 to 2017 to examine the prediction performance of Z-SVR and compared its performance with those of other widely used machine learning methods. The results show that Z-SVR outperformed the other machine learning methods and can further enhance the accuracy of casualty prediction.

P4622Prediction of in-hospital bleeding for AMI patients undergoing PCI using machine learning method

2019 ◽  
Vol 40 (Supplement_1) ◽  
Author(s):  
X Y Zhao ◽  
J G Yang ◽  
T G Chen ◽  
J M Wang ◽  
X Li ◽  
...  
Keyword(s):  
Machine Learning ◽  
Myocardial Infarction ◽  
Acute Myocardial Infarction ◽  
Prediction Model ◽  
Clinical Decision Making ◽  
Learning Algorithm ◽  
Area Under The Curve ◽  
Learning Methods ◽  
Machine Learning Methods ◽  
Validation Set

Abstract Background Prediction of in-hospital bleeding is critical for clinical decision making for acute myocardial infarction (AMI) patients undergoing percutaneous coronary intervention (PCI). Machine learning methods can automatically select the combination of the important features and learn their underlying relationship with the outcome. Objective We aim to evaluate the predictive value of machine learning methods to predict in-hospital bleeding for AMI patients. Methods We used data from the multicenter China Acute Myocardial Infarction (CAMI) registry. We randomly partitioned the cohort into derivation set (75%) and validation set (25%). Using data from the derivation set, we applied a state-of-art machine learning algorithm, XGBoost, to automatically select features from 106 candidate variables and train a risk prediction model to predict in-hospital bleeding (BARC 3, 5 definition). Results 16736 AMI patients who underwent PCI were consecutively included in the analysis, while 70 (0.42%) patients had in-hospital bleeding followed the BARC 3,5 definition of bleeding. Fifty-nine features were automatically selected from the candidate features and were used to construct the prediction model. The area under the curve (AUC) of the XGBoost model was 0.816 (95% CI: 0.745–0.887) on the validation set, while AUC of the CRUSADE risk score was 0.723 (95% CI: 0.619–0.828). Relative contribution of the 12 most important features Feature Relative Importance Direct bilirubin 0.078 Heart rate 0.077 CKMB 0.076 Creatinine 0.064 GPT 0.052 Age 0.048 SBP 0.036 TG 0.035 Glucose 0.035 HCT 0.031 Total bilirubin 0.030 Neutrophil 0.030 ROC of the XGBoost model and CRUSADE Conclusion The XGBoost model derived from the CAMI cohort accurately predicts in-hospital bleeding among Chinese AMI patients undergoing PCI. Acknowledgement/Funding the CAMS innovation Fund for Medical Sciences (CIFMS) (2016-12M-1-009); the Twelfth Five-year Planning Project of China (2011BAI11B02)

Sign in / Sign up

Export Citation Format

Share Document