Robustness of Support Vector Regression and Random Forest Models: A Simulation Study

Classical statistics are usually based on parametric models, where the performance depends heavily on assumptions and is not robust in the presence of outliers in the data. Due to the COVID-19 pandemic, our daily lives have changed significantly, including slowing economic growth. These extreme changes can manifest as an outlier in time series studies and adversely affect the results of data analysis. Many classical methods of official statistics are prone to outliers. In this work, we evaluate machine learning methods: Support Vector Regression (SVR) and Random Forest (RF) and compare it with ARIMA to determine the robustness through simulation studies. Robustness is measured by the sensitivity of the SVR and Random Forest hyperparameter and the model's error in the presence of outliers. Simulations show that more outliers lead to higher RMSE values, and conversely, more samples lead to lower RMSE values. The type of outliers significantly impacts the RMSE value of the ARIMA model, where additional outliers (AO) have a worse impact than temporary change (TC). Consecutive outliers produce a smaller RMSE mean than non-consecutive outliers. Based on the sensitivity of hyperparameters, SVR and Random Forest models are relatively robust to the presence of outliers in the data. Based on the simulation results of 100 iterations, we find that SVR is more robust than ARIMA and Random Forest in modeling time series data with outliers. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

Estimating Medical Insurance Cost using Linear Regression with HyperParameterization, Decision Tree and Random Forest Models

The total health expenditure refers to the total public and private funds spent on health services and amenities, medical and surgical bills and all other healthcare facilities provided. The financing for health is of great significance and plays a crucial role in health systems. To enhance the productivity of human capital, the efficiency and delivery of healthcare services must be uplifted. Reports have shown that from year 2000 to 2018 there had been a gradual escalation of global health expenses and is standing on around 10% of the total GDP of the world. Out of pocket expenses are also high in least developed nations that have lower per capita income. Even though the World Health Organization (WHO) sanctions loans to these countries, these nations are bound to use the money on industrialization only and not their healthcare, education and public welfare sectors. With inflation, the expenses of first-rate healthcare are also rising which makes it fundamental to have health and life insurance plans. Health insurance schemes insured around 514 million people in India in the year 2021, most of which were covered under government schemes only. Since the advent of COVID-19 people have realized the need for having a insurance plan. Most of the companies that are based on the health insurance sector use predictive modelling to improve their services and business process. Machine Learning (ML) algorithms are used to train a model and provide insurance costs estimations. Past data is searched for any pattern or trend in the behaviour history of consumers and then future estimations are evaluated. The proposed project is comprised of different regression models like Linear regression with hyperparameterization , regressors like Decision Tree and Random forest to estimate the approximate insurance expenditure. © 2023 IEEE.

Decision Tree and Random Forest Models for Internal Audit Effectiveness Factors Evaluation

this paper investigates the impact of auditors' competency, independency, member's size, and digital transformation as factors that enhance the effectiveness of internal audit in Bahraini Context during and after Covid-19 outbreak. In this quantitative study, this paper used a survey questionnaire. The survey used to collect data for a sample of 50 respondents from both public and private organization in Bahrain to measure the study variables among auditors. Findings, there is a strong relationship between internal audit effectiveness and management support, digital transformation, and IA member's size. While independency, competency, and auditors' experience positively affect the internal audit effectiveness. The results of this paper will encourage decision makers of such organizations to focus on the issue of that enhance the effectiveness of internal audit after pandemic outbreaks and will strengthen the internal audit capacity of these organizations. This Paper offer academic contributions to existing research of internal audit effectiveness in developing context. The decision trees and Random Forest approaches used in this paper will allow future researchers to reclassify the factors that enhance the effectiveness of internal audit. © 2023 IEEE.

Application of machine learning approaches to predict the impact of ambient air pollution on outpatient visits for acute respiratory infections

With a remarkable increase in industrialization among fast-developing countries, air pollution is rising at an alarming rate and has become a public health concern. The study aims to examine the effect of air pollution on patient's hospital visits for respiratory diseases, particularly Acute Respiratory Infections (ARI). Outpatient hospital visits, air pollution and meteorological parameters were collected from March 2018 to October 2021. Eight machine learning algorithms (Random Forest model, K-Nearest Neighbors regression model, Linear regression model, LASSO regression model, Decision Tree Regressor, Support Vector Regression, X.G. Boost and Deep Neural Network with 5-layers) were applied for the analysis of daily air pollutants and outpatient visits for ARI. The evaluation was done by using 5-cross-fold confirmations. The data was randomly divided into test and training data sets at a scale of 1:2, respectively. Results show that among the studied eight machine learning models, the Random Forest model has given the best performance with R2 = 0.606, 0.608 without lag and 1-day lag respectively on ARI patients and R2 = 0.872, 0.871 without lag and 1-day lag respectively on total patients. All eight models did not perform well with the lag effect on the ARI patient dataset but performed better on the total patient dataset. Thus, the study did not find any significant association between ARI patients and ambient air pollution due to the intermittent availability of data during the COVID-19 period. This study gives insight into developing machine learning programs for risk prediction that can be used to predict analytics for several other diseases apart from ARI, such as heart disease and other respiratory diseases. © 2022 Elsevier B.V.

Reevaluating Synthesizing Sentiment Analysis on COVID-19 Fake News Detection using Spark Dataframe

Some research uses the random forest model and sentiment analysis to detect COVID-19 fake news. However, there is still a research opportunity to apply the method to Indonesian Tweets and reevaluate the feature's performance. Our research aims to reevaluate synthesizing the sentiment analysis feature on detecting COVID-19 fake news on Indonesian Tweets by using the Spark Dataframe. We divide the stages of machine learning development into several steps, including collecting data using Tweepy and then applying sentiment polarity scores using Apache Spark. We apply random forest to classify fake news using the Spark MLlib. Further, we use model evaluation calculation through the level of Accuracy, Recall, Precision, and F1. The results show that applying the sentiment polarity calculation to our Tweet dataset labels 148 Tweets with positive sentiments, 118 Tweets with negative sentiments, and 99 Tweets with neutral sentiments. The Pearson correlation coefficient (PCC) feature score of Sentiment equals 0.056 and ranks fifth in the top feature correlation scores list. According to the experimental findings, the random forest model produces Accuracy = 0.787 for both models with sentiment analysis and without sentiment analysis. Which indicates that sentiment analysis provides no significance in the prediction model. © 2022 IEEE.

Vaccine Supply Optimization and Forecasting using Random Forest and ARIMA Models

In order to tackle the Corona Virus Disease, it took a considerable amount of time for the governments to come up with effective and efficient vaccines. After the vaccines were developed, the next challenge was to supply the vaccines to various designated centers based on demographics, population distribution, and other factors. The whole system for vaccine supply played a vital role during the COVID-19 pandemic. We also saw a lot of haphazard and mismanagement in some places especially when the cases per day surged high, as people weren't prepared for such a situation. Now that we have got enough data, we can use it to optimize the vaccine supply across various Covid Vaccination Centers and be prepared for any such circumstances in the future. In this paper, we have proposed a two-step approach where considering the past supply and wastage data we performed a classification task that indicates whether doses are to get wasted at a given center. If yes, we then perform demand forecasting based on the number of administered doses so that the wastage can be reduced, and supply can be optimized. © 2022 IEEE.

Addressing Cancer Readmission Prediction Model Drift: A Case Study

The cancer readmission prediction model classifies patients as high-risk or low-risk for readmittance. Consequently, intervention strategies focus on high-risk patients. Nevertheless, the performance of machine learning models generally degrades over time due to changes in the environment that violates models' assumptions, which include statistical data changes and process changes. This research introduces a framework that improves the sensitivity of the cancer readmission prediction model by identifying new features of cancer readmission, such as Diabetes and Anti-Nausea, which potentially cause the model's sensitivity to drift. The proposed model considers these 20 new factors with the 35 original factors that use the most recent dataset to predict cancer readmissions. Recursive feature elimination was used to identify key features. Some of the most popular classification algorithms, which include logistic regression and adaptive boosting, were used to retrain and classify cancer readmissions. The best algorithm was validated on a new dataset that was collected over 11 months, which covered three different waves of Covid-19. The results suggested K-Nearest Neighbors (KNN) algorithm performs the best among all eight studied algorithms. The KNN model incorporated new dominant features that did not exist in the original Random Forest (RF) model. The KNN model has an improvement of 8.05% in sensitivity compared to the RF model. The presence of Covid-19 does not have a significant impact on the performance of the KNN model. The suggested framework identifies potential admitted patients for intervention actions, helps reduce cancer readmission rates, costs, and improves the quality of care for cancer patients. © 2022 IISE Annual Conference and Expo 2022. All rights reserved.

A Machine Learning Model for Effective Consumer Behaviour Prediction

Effective consumer behavior prediction can play a crucial role in online marketing, especially in the COVID19 scenario. In this work, we have analyzed consumer behavior to understand consumer needs and predict future requirements. For the same, we have applied the machine learning models on an amazon dataset collected from Kaggle. The dataset consists of reviewers' comments, ratings, many other parameters for the product. The model's outcome indicates that the proposed Random Forest model performs exceptionally well, and its Accuracy is approx. 98.73%. A comparative study has been done to show the efficacy of the work, and it has been observed that the performance of the proposed model is quite remarkable, and it can be a competent model for effective consumer behavior prediction. © 2021 IEEE.

COVID-19 Vaccine Distribution Policy Design with Reinforcement Learning

COVID-19 has become a global crisis and the vaccine has been seen as an effective approach to stop the epidemic spread. However, the resources for distributing and allocating different types of vaccines are limited and we need a better vaccine distribution policy design to prevent the spread of COVID-19 more efficiently. In this study, a pipeline of combing a random forest model and a DQN model is proposed. The random forest model is built to predict the daily new confirmed cases with the vaccine data as the inputs. And the DQN model is built to design the daily allocation ratio of three types of vaccines, with the aim to minimize the new confirmed cases. The experimental results based on the real-world datasets collected in San Diego validate the effectiveness of the proposed pipeline. © 2021 ACM.