• The Korean Journal of Applied Statistics
• /
• v.32 no.3
• /
• pp.391-404
• /
• 2019

Customers' fixed characteristics have often been used to predict customer behavior. It has recently become possible to track customer web logs as customer activities move from offline to online. It has become possible to collect large amounts of web log data; however, the researchers only focused on organizing the log data or describing the technical characteristics. In this study, we predict the decision-making time until each customer makes the first reservation, using Airbnb customer data provided by the Kaggle website. This data set includes basic customer information such as gender, age, and web logs. We use various methodologies to find the optimal model and compare prediction errors for cases with web log data and without it. We consider six models such as Lasso, SVM, Random Forest, and XGBoost to explore the effectiveness of the web log data. As a result, we choose Random Forest as our optimal model with a misclassification rate of about 20%. In addition, we confirm that using web log data in our study doubles the prediction accuracy in predicting customer behavior compared to not using it.

• Journal of Digital Policy
• /
• v.1 no.2
• /
• pp.61-69
• /
• 2022

The paper is a study to develop and verify a blood glucose level prediction model based on biosignals obtained from photoplethysmography (PPG) sensors, ICT technology and data. Blood glucose prediction used the MLP architecture of machine learning. The input layer of the machine learning model consists of 10 input nodes and 5 hidden layers: heart rate, heart rate variability, age, gender, VLF, LF, HF, SDNN, RMSSD, and PNN50. The results of the predictive model are MSE=0.0724, MAE=1.1022 and RMSE=1.0285, and the coefficient of determination (R²) is 0.9985. A blood glucose prediction model using bio-signal data collected from digital devices and machine learning was established and verified. If research to standardize and increase accuracy of machine learning datasets for various digital devices continues, it could be an alternative method for individual blood glucose management.

• Journal of Sasang Constitutional Medicine
• /
• v.33 no.4
• /
• pp.1-9
• /
• 2021

Objectives The purpose of this study was to check whether the classification of the individual's Body Mass Index (BMI) could be predicted by analyzing the voice data constructed at the Korean medicine data center (KDC) using machine learning. Methods In this study, we proposed a convolutional neural network (CNN)-based BMI classification model. The subjects of this study were Korean adults who had completed voice recording and BMI measurement in 2006-2015 among the data established at the Korean Medicine Data Center. Among them, 2,825 data were used for training to build the model, and 566 data were used to assess the performance of the model. As an input feature of CNN, Mel-frequency cepstral coefficient (MFCC) extracted from vowel utterances was used. A model was constructed to predict a total of four groups according to gender and BMI criteria: overweight male, normal male, overweight female, and normal female. Results & Conclusions Performance evaluation was conducted using F1-score and Accuracy. As a result of the prediction for four groups, The average accuracy was 0.6016, and the average F1-score was 0.5922. Although it showed good performance in gender discrimination, it is judged that performance improvement through follow-up studies is necessary for distinguishing BMI within gender. As research on deep learning is active, performance improvement is expected through future research.

• Journal of Korean Biological Nursing Science
• /
• v.26 no.1
• /
• pp.26-40
• /
• 2024

Purpose: The aim of this study was to predict the subgroups vulnerable to poorer health-related quality of life (HRQoL) according to gender in older adults. Methods: Data from 5,553 Koreans aged 65 or older were extracted from the Korea National Health and Nutrition Examination Survey. HRQoL was assessed using the EQ-5D tool. Complex sample analysis and decision-tree analysis were conducted using SPSS for Windows version 27.0. Results: The mean scores of the EQ-5D index were 0.93 ± 0.00 in men and 0.88 ± 0.00 in women. In men, poorer HRQoL groups were identified with seven different pathways, which were categorized based on participants' characteristics, such as restriction of activity, perceived health status, muscle exercise, age, relative hand grip strength, suicidal ideation, the number of chronic diseases, body mass index, and income status. Restriction of activity was the most significant predictor of poorer HRQoL in elderly men. In women, the poorer HRQoL groups were identified with nine different pathways, which were categorized based on participants' characteristics, such as perceived health status, restriction of activity, age, education, unmet medical service needs, anemia, body mass index, relative hand grip, and aerobic exercise. Perceived health status was the most significant predictor of poorer HRQoL in elderly women. Conclusion: This study presents a predictive model of HRQoL in older adults according to gender and can be used to detect individuals at risk of poorer HRQoL.

• Journal of Preventive Medicine and Public Health
• /
• v.32 no.1
• /
• pp.60-64
• /
• 1999

Objectives. Changes in lung function are frequently used as biological markers to assess the health effects of criteria air pollutants. We tried to formulate the prediction models of pulmonary functions based on height, weight, age and gender, especially for children aged 12 years who are commonly selected for the study of health effects of the air pollution. Methods. The target pulmonary function parameters were forced vital capacity(FVC) and forced expiratory volume in one second(FEV1). Two hundreds and fifity-eight male and 301 female 12-year old children were included in the analysis after excluding unsatisfactory tests to the criteria recommended by American Thoracic Sosiety and excluding more or less than 20% predicted value by previous prediction equations. The weight prediction equation using height as a independent variable was calculated, and then the difference of observed weight and predicted weight (i.e. residual) was used as the independent variable of pulmonary function prediction equations with height. Results. The prediction equations of FVC and FEV1 for male are FVC(ml) = $50.84{\times}height(cm)+7.06{\times}weight$ residual 4838.86, FEV1(ml) = $43.57{\times}height(cm)+3.16{\times}weight$ residual - 4156.66, respectively. The prediction equations of FVC and FEV1 for female are FVC(ml) = $42.57{\times}height(cm)+12.50{\times}weight$ residual - 3862.39, FEV1(ml) = $36.29{\times}height(cm)+7.74{\times}weight$ residual - 3200.94, respectively.

• The Journal of the Convergence on Culture Technology
• /
• v.8 no.5
• /
• pp.207-218
• /
• 2022

This study, to confirm the mediating effect of knee pain on cognitive functions and depression in older adults, and as an interdisciplinary research between the physical and psychological mechanisms, confirmed the identifying group differences by gender, age, and educational attainment of older adults, and aimed to research the improvement of cognitive functions, which is a main factor of dementia's risk prediction. The analysis data was from the 8th Korean Longitudinal Study of Ageing (KLoSA) in 2020, and the research model was verified using Process macro and model #4. The main analysis results are as follows. First, depression partially mediation effect of knee pain on cognitive functions. Second, the mediation effect of depression by gender was significant, but the direct effect in the male older adults group was twice that in the female older adults; the indirect effect did not vary significantly based on gender. Third, the mediating effect of depression by age was relatively greater in the old-old aged group than in the young-old aged one. Fourth, as for the mediating effect of depression according to the classification of educational attainment, the mediating effect was not significant in the group with a college degree or higher education but was significant in the remaining three sub-groups. Based on the results, this study makes implications for the need for active intervention strategies to improve cognitive functions, focusing on group differences by gender, age, and educational attainment in the management of knee pain and depression.

• Health Policy and Management
• /
• v.33 no.4
• /
• pp.450-456
• /
• 2023

Background: This study aims to develop a "Predictive Model for the Possibility of Collection Delinquent Health Insurance Contributions" for the National Health Insurance Service to enhance administrative efficiency in protecting and collecting contributions from livelihood-type defaulters. Additionally, it aims to establish customized collection management strategies based on individuals' ability to pay health insurance contributions. Methods: Firstly, to develop the "Predictive Model for the Possibility of Collection Delinquent Health Insurance Contributions," a series of processes including (1) analysis of defaulter characteristics, (2) model estimation and performance evaluation, and (3) model derivation will be conducted. Secondly, using the predictions from the model, individuals will be categorized into four types based on their payment ability and livelihood status, and collection strategies will be provided for each type. Results: Firstly, the regression equation of the prediction model is as follows: phat = exp (0.4729 + 0.0392 × gender + 0.00894 × age + 0.000563 × total income - 0.2849 × low-income type enrollee - 0.2271 × delinquency frequency + 0.9714 × delinquency action + 0.0851 × reduction) / [1 + exp (0.4729 + 0.0392 × gender + 0.00894 × age + 0.000563 × total income - 0.2849 × low-income type enrollee - 0.2271 × delinquency frequency + 0.9714 × delinquency action + 0.0851 × reduction)]. The prediction performance is an accuracy of 86.0%, sensitivity of 87.0%, and specificity of 84.8%. Secondly, individuals were categorized into four types based on livelihood status and payment ability. Particularly, the "support needed group," which comprises those with low payment ability and low-income type enrollee, suggests enhancing contribution relief and support policies. On the other hand, the "high-risk group," which comprises those without livelihood type and low payment ability, suggests implementing stricter default handling to improve collection rates. Conclusion: Upon examining the regression equation of the prediction model, it is evident that individuals with lower income levels and a history of past defaults have a lower probability of payment. This implies that defaults occur among those without the ability to bear the burden of health insurance contributions, leading to long-term defaults. Social insurance operates on the principles of mandatory participation and burden based on the ability to pay. Therefore, it is necessary to develop policies that consider individuals' ability to pay, such as transitioning livelihood-type defaulters to medical assistance or reducing insurance contribution burdens.

• Physical Therapy Korea
• /
• v.18 no.3
• /
• pp.76-84
• /
• 2011

The purpose of this study was to investigate the correlations between pain intensity, physical impairments, disability, and psychological factors according to the difference in duration of low back pain. This study was a cross-sectional survey of 102 participants with low back pain, divided into two groups equal in number: The first group consisted of patients with acute and subacute low back pain, while the second group consisted of patients suffering from chronic low back pain. The results showed that gender, age, pain intensity, physical impairment, disability and Fear-Avoidance Beliefs (FABs) for work activities were not significantly different between two groups. FABs for physical activities of the first group were significantly more prevalent than in the second group. More than moderate correlations were found between pain intensity, physical impairment, and disability in the first group. Less than moderate correlations were found between pain intensity, physical impairment, disability, FABs, and depression in the second group. These findings suggest that we must consider psychological factors in the treatment of patients with chronic low back pain. Regression analyses revealed that pain intensity and FABs for work activities significantly contributed to the prediction of disability in the first group. Also, pain intensity and FABs for physical activities significantly contributed to the prediction of disability in the second group. Pain intensity was most important predictor of disability in two groups.

• Journal of Korean Academy of Fundamentals of Nursing
• /
• v.18 no.3
• /
• pp.317-327
• /
• 2011

Purpose: The purpose of this study was to contribute data toward prevention from advancing dementia and also prevention of deterioration in cognitive impairment by constructing an optimal prediction model and verifying factors influencing cognitive impairment in elders with dementia who reside at home. Methods: The participants in this study were 351 elders who were registered at dementia day care centers in 11 regions of Metropolitan Incheon. Collected data were analyzed using SPSS Statistics 17.0 and SAS 9.1. Bootstrap method using the Clementine program 12.0 was applied to build an optimum prediction model. Results: Gender and education (general characteristics), alcohol, urinary/fecal incontinence, exercise, weight, and ADL (state of health), and depression (psychological state) were found to have an affect on cognitive impairment in these elders. Conclusion: Study results indicate nine key factors that affect cognitive impairment of elders with dementia who reside at home and that could be useful in prevention and management nursing plans. These factors could also be used to expand the role of nurses who are working in community day care centers, and can be applied in the development and provision of various programs to aid retention and improve cognitive function as well as preventing deterioration of cognition.

• International Journal of Computer Science & Network Security
• /
• v.21 no.3
• /
• pp.103-111
• /
• 2021

Obesity affects individuals of all gender and ages worldwide; consequently, several studies have performed great works to define factors causing it. This study develops an effective method to trace obesity levels based on supervised data mining techniques such as Random Forest and Multi-Layer Perception (MLP), so as to tackle this universal epidemic. Notably, the dataset was from countries like Mexico, Peru, and Colombia in the 14- 61year age group, with varying eating habits and physical conditions. The data includes 2111 instances and 17 attributes labelled using NObesity, which facilitates categorization of data using Overweight Levels l I and II, Insufficient Weight, Normal Weight, as well as Obesity Type I to III. This study found that the highest accuracy was achieved by Random Forest algorithm in comparison to the MLP algorithm, with an overall classification rate of 96.7%.