SAS Enterprise Miner Research Articles

Health Insurance Premium Pricing Using Machine Learning Methods

Health insurance is important alongside life insurance products. This product’s subscription is gradually rising and has become one of the public’s main considerations due to their awareness on medical and surgical costs. This study is aimed at (i) identifying whether the independent factors are important in predicting the dependent variable, and (ii) to determine which model (logistic regression model/decision trees/neural networks) is the best model to be utilised. The SAS Enterprise Miner (E-Miner) was used to analyse the data and to select the best model. At the end of this study, the measurements like the Maximum Absolute Error (MAE), Average Squared Error (ASE), Root Average Squared Error (RASE), and Sum Squared Error (SSE) in the decision tree model indicated the lowest errors and led to the selection of the decision tree as the best model to be used in health insurance premium pricing.

A Predictive Model of Ischemic Heart Disease in Middle-Aged and Older Women Using Data Mining Technique.

This study was conducted to identify ischemic heart disease-related factors and vulnerable groups in Korean middle-aged and older women using data from the Korea National Health and Nutrition Examination Survey (KNHANES). Among the 24,229 people who participated in the 2017-2019 survey, 7249 middle-aged women aged 40 and over were included in the final analysis. The data were analyzed using IBM SPSS and SAS Enterprise Miner by chi-squared analysis, logistic regression analysis, and decision tree analysis. The prevalence of ischemic heart disease in the study results was 2.77%, including those diagnosed with myocardial infarction or angina. The factors associated with ischemic heart disease in middle-aged and older women were identified as age, family history, hypertension, dyslipidemia, stroke, arthritis, and depression. The group most vulnerable to ischemic heart disease included women who had hypertension, a family history of ischemic heart disease, and were menopausal. Based on these results, effective management should be achieved by applying customized medical services and health management services for each relevant factor in consideration of the characteristics of the groups with potential risks. This study can be used as basic data that can be helpful in national policy decision making for the management of chronic diseases.

Perceived injustice and brand love: the effectiveness of sympathetic vs empathetic responses to address consumer complaints of unjust specific service encounters

PurposeNegative online consumer reviews represent different forms of injustice. The effect of different types of injustice experienced in a service encounter on a brand is unknown. This study aims to investigate the effect and cause of different forms of injustice on brand love. It also explores which type of responses are more effective to mitigate their damaging effect.Design/methodology/approachOne text mining, using SAS enterprise miner, and three experimental studies were conducted. ANOVA and mediation and moderation analyses were conducted to test the hypotheses.FindingsNegative reviews specific to procedural injustice are more damaging than reviews specific to distributive or interactional injustice experienced in a service encounter. The underlying reason behind this differential effect is that perceived procedural injustice influences consumers more to punish the brand, resulting in a greater negative effect on brand love. To counter the damage, a sympathetic, rather than empathetic, brand response is more effective.Originality/valueThis study contributes to justice theory and brand love literature by providing evidence that procedural injustice triggers the highest level of willingness to punish and thus the lowest level of brand love. Consequently, willingness to punish, rather than emotion, is found to be the underlying reason behind procedural injustice having the strongest negative effect on brand love.

Text mining the mission statements of the most ethical companies

PurposeThis paper explores and examines the mission statements of the most ethical companies across the globe in terms of their main purposes, values, goals, and objective, and what they say about their vision and goals.Design/methodology/approachThis study is based on the data published by the Ethisphere Institute, the global leader in defining and advancing the standards of ethical business practices. Having compiled the mission statements into a text file, the authors conducted text mining using a commercially available text mining tool SAS Enterprise Miner to survey if the most ethical companies have valued the same vision and mission such as social responsibility and ethics.FindingsA review of their mission statements indicated that some of the most ethical companies surveyed in this study such as 3M and Voya strive to be “socially responsible and ethical,” support their “societies” and respect and protect the “nature,” “planet” and “environment.” The world's most ethical companies that stress these weighted terms in their mission statements may do so to show their commitment by being socially responsible and ethical, and delivering sustainable business solutions to their customers.Originality/valueThis study provides a systematic and comprehensive exploration of mission statements of the most ethical companies in an attempt to identify patterns of differences and similarities within these statements.

Defining characteristics of the most innovative companies

In this study, a text analytics approach was employed to analyse the mission statements of the most innovative companies to help uncover emerging trends and unlock the value of unstructured textual data. The tools that were used to conduct this analysis were R and SAS Enterprise Miner. The dataset that is used for this analysis contains the mission statements of the top 50 most innovative companies ranked by Boston Consulting Group. Valuable information was extracted by classifying, clustering, and visualising the most frequently appearing and significant terms found across the mission statements of the most innovative companies.

Negative online reviews, brand equity and emotional contagion

PurposeDespite the growing consensus that consumers extensively use online reviews and that negative reviews can significantly damage brand equity, it remains uncertain whether negative online reviews that focus on different aspects of a service have a similar or differential effect on brand equity. This study aims to fill this gap and explores the mediating role of emotional contagion and what kind of response helps better deter their negative effect.Design/methodology/approachThis research is conducted through a one-panel study and three experimental studies. SAS enterprise miner is used for text mining analysis and Analysis of variance (ANOVA) and Process macro models are used to analyze the experimental data.FindingsNegative reviews related to the tangibility, responsiveness and empathy dimensions have a more detrimental effect on brand equity than negative reviews related to the assurance and reliability dimensions. The results also provide evidence that emotional contagion is more prevalent when consumers read reviews that are specific to the empathy and responsiveness dimensions. Finally, accommodative responses from the service provider are more effective in deterring the effect of a negative online review on brand equity.Research limitations/implicationsThe generalizability of this study is limited to the restaurant and hotel industry.Practical implicationsThe findings will also help the brand manager in understanding the comparative effect of service quality-specific negative reviews on brand equity and also the type of responses that brand managers should give to negative reviews.Originality/valueDespite online reviews receiving increased attention in academic research, Service quality (SERVQUAL) dimension-specific reviews have not been studied until now. This study contributes to the service quality-related literature by providing evidence that not all negative online reviews related to different Service quality (SERVQUAL) dimensions equally affect brand equity.

Trade Heterogeneity in the EU: Insights from the Emergence of COVID-19 Using Time Series Clustering

Objective: The objective of the paper is to analyse segmentation of EU-27 countries based on quarterly growth rates of exports and imports by using time series clustering.Research Design & Methods: We applied a time series clustering algorithm using TS nodes in SAS Enterprise Miner. To analyse the impact of the pandemic, we considered clusters based on export and import growth rates for two time periods, pre-emergence and post-COVID-19 emergence.Findings: We find that grouping based on export and import growth rates vary for EU-27 countries. Also, clustering results change significantly for post-COVID-19 emergence compared to pre-COVID-19 emergence. Cyprus emerged as an exception based on export growth rates, while Malta came out as an outlier based on the segmentation of its import growth rates.Implications / Recommendations: The impact and severity of COVID-19 has varied across EU countries, which have shown a varied impact in their trade patterns characterised by growth rates of exports and imports. The clustering analysis presented in the paper helps to explain similarities and differences in trade patterns of EU members during the COVID-19 pandemic to effectively implement and harmonise EU specific trade policies to member countries.Contribution: The study contributes to the literature on EU trade by providing an approach to analysing EU-27 segments using time series clustering analysis. It also enhances the growing literature on the impact of the pandemic on international trade by separating clustering analysis for the COVID-19 period and investigating the drivers for the segmentation.

Predicting Breast Cancer Survivability

this study concentrates on Predicting Breast Cancer Survivability using data mining, and comparing between three main predictive modeling tools. Precisely, we used three popular data mining methods: two from machine learning (artificial neural network and decision trees) and one from statistics (logistic regression), and aimed to choose the best model through the efficiency of each model and with the most effective variables to these models and the most common important predictor. We defined the three main modeling aims and uses by demonstrating the purpose of the modeling. By using data mining, we can begin to characterize and describe trends and patterns that reside in data and information. The preprocessed data set contents were of 87 variables and the total of the records are 457,389; which became 93 variables and 90308 records for each variable, and these dataset were from the SEER database. We have achieved more than three data mining techniques and we have investigated all the data mining techniques and finally we find the best thing to do is to focus about these data mining techniques which are Artificial Neural Network, Decision Trees and Logistic Regression by using SAS Enterprise Miner 5.2 which is in our view of point is the suitable system to use according to the facilities and the results given to us. Several experiments have been conducted using these algorithms. The achieved prediction implementations are Comparison-based techniques. However, we have found out that the neural network has a much better performance than the other two techniques. Finally, we can say that the model we chose has the highest accuracy which specialists in the breast cancer field can use and depend on.

МЕТОДИКА СКОРИНГОВОГО АНАЛИЗА В ИССЛЕДОВАНИИ ДИНАМИКИ РАЗВИТИЯ ИННОВАЦИОННЫХ СТАРТАПОВ НА ПРИМЕРЕ СФЕРЫ ОКАЗАНИЯ УСЛУГ

In the article the model of ability to live of innovative startups in sphere of rendering of services on the basis of an administrative configuration scoring the models which advantages is use of product SAS Enterprise Miner 5.1 and product introduction SAS Servise Scoring is developed.

A content analysis of top-ranked universities’ mission statements from five global regions

In general, a mission statement reflects the context in which a higher education institution operates. It’s also a direction setting statement indicating the direction required for effective strategic and operational decisions. This study examines how the highest ranked universities in Europe, Asia, Latin America, Africa, and North America compare in terms of their mission statements and how they communicate to their stakeholders and to the public about what their main purposes, values, goals, and objectives are. Having downloaded each university’s mission statement from the Internet, we then classified them into five main categories of Asian, African, European, Latin American, and North American universities and analyzed them separately using the text mining and analysis tools available in SAS Enterprise Miner, a powerful commercial business analytics program. Results show that although we found similar terms and themes in the mission statements across the five different regions, there is great variability in the mission statements with respect to how they communicate their core values, main purposes, and value propositions.

Text Mining Method for Studying Medication Administration Incidents and Nurse-Staffing Contributing Factors: A Pilot Study.

Incident reporting systems are being implemented globally, thus increasing the profile and prevalence of incidents, but the analysis of free-text descriptions remains largely hidden. The aims of the study were to explore the extent to which incident reports recorded staffing issues as contributors to medication administration incidents. Incident reports related to medication administration (N = 1012) were collected from two hospitals in Finland between January 1, 2013, and December 31, 2014. The SAS Enterprise Miner 13.2 and its Text Miner tool were used to excavate terms and descriptors and to uncover themes and concepts in the free-text descriptions of incidents with (n = 194) and without (n = 818) nurse staffing-related contributing factors. Text mining included (1) text parsing, (2) text filtering, and (3) modeling text clusters and text topics. The term "rush/hurry" was the sixth most common term used in incidents where nurse-staffing was identified as a contributing factor. Nurse-staffing factors, however, were not pronounced in clusters or in text topics of either data set. Text mining offers the opportunity to analyze large free-text mass and holds promise for providing insight into the antecedents of medication administration incidents.

Personal bankruptcy prediction using decision tree model

Purpose Personal bankruptcy is on the rise in Malaysia. The Insolvency Department of Malaysia reported that personal bankruptcy has increased since 2007, and the total accumulated personal bankruptcy cases stood at 131,282 in 2014. This is indeed an alarming issue because the increasing number of personal bankruptcy cases will have a negative impact on the Malaysian economy, as well as on the society. From the aspect of individual’s personal economy, bankruptcy minimizes their chances of securing a job. Apart from that, their account will be frozen, lost control on their assets and properties and not allowed to start any business nor be a part of any company’s management. Bankrupts also will be denied from any loan application, restricted from travelling overseas and cannot act as a guarantor. This paper aims to investigate this problem by developing the personal bankruptcy prediction model using the decision tree technique. Design/methodology/approach In this paper, bankrupt is defined as terminated members who failed to settle their loans. The sample comprised of 24,546 cases with 17 per cent settled cases and 83 per cent terminated cases. The data included a dependent variable, i.e. bankruptcy status (Y = 1(bankrupt), Y = 0 (non-bankrupt)) and 12 predictors. SAS Enterprise Miner 14.1 software was used to develop the decision tree model. Findings Upon completion, this study succeeds to come out with the profiles of bankrupts, reliable personal bankruptcy scoring model and significant variables of personal bankruptcy. Practical implications This decision tree model is possible for patent and income generation. Financial institutions are able to use this model for potential borrowers to predict their tendency toward personal bankruptcy. Social implications Create awareness to society on significant variables of personal bankruptcy so that they can avoid being a bankrupt. Originality/value This decision tree model is able to facilitate and assist financial institutions in evaluating and assessing their potential borrower. It helps to identify potential defaulting borrowers. It also can assist financial institutions in implementing the right strategies to avoid defaulting borrowers.

Text Mining And Sentiment Analysis On Reviews Of Proton Cars In Malaysia

Social networks play an important role in commercial products, and thus knowing the emotions and opinions of users is very useful in improving services, sales, business and marketing strategies. This paper illustrates text mining approach and sentiment analysis to gain valuable insights into consumer perceptions towards Proton cars in Malaysia. In the first case study, a total of 5000 comments posted on Paultan’s Automotive News Facebook page were retrieved and analyzed using R, a statistical open source programming software. The final dataset consists of 277 posts with 2964 comments. Data cleaning was carried out to remove non-English word. A word cloud was then generated and the frequently mentioned words are “China”, “like”, “price” and “good” and ‘SUV’. In the second case study, we perform sentiment analysis on a total of 5330 comments using SAS Enterprise Miner 14.1. Out of the 60 documents, 39 (65%) posts were positive comments from the public while 21 (35%) posts were negative comments. The frequently mentioned words in the positive documents are “look good”, “buy”, “better” and “nice”. Chi-Square and Information Gain were compared in selecting the meaningful terms of features. The selected features were used to evaluate the performance of Support Vector Machine (SVM) in classifying the posts as positive or negative. Five-fold cross validation results showed that SVM using linear kernel function has the highest accuracy (73.3%), sensitivity (76.7%) and F-measure (0.805).

Group regularization for zero-inflated poisson regression models with an application to insurance ratemaking

ABSTRACTZero-inflated count models have received considerable amount of attention in recent years, fuelled by their widespread applications in many scientific disciplines. In this paper, we consider the problem of selecting grouped variables in zero-inflated Poisson (ZIP) models via group bridge regularization. The ZIP mixture likelihood with a group-wise penalty on the coefficients is formulated using least squares approximation and then the parameters involved in the penalized likelihood are estimated by an efficient group descent algorithm. We examine the effectiveness of our modeling procedure through extensive Monte Carlo simulations. An auto insurance claim dataset from the SAS Enterprise Miner database is analyzed for illustrative purposes. Finally, we derive theoretical properties of the proposed group variable selection procedure under certain regularity conditions. The open source software implementation of this method is publicly available at https://github.com/himelmallick/Gooogle.

Quality improvement using new technology to assess and reduce FDG PET/CT radiotracer infiltrations.

313 Background: The reported aggregated PET/CT 18F-FDG radiotracer injection infiltration rate from 3 centers in 6 studies is 15.2% (425 infiltrations/2804 patients). Infiltrations can negatively affect cancer patient staging, therapy assessment, treatment planning, and can lead to unnecessary invasive procedures and patient radiation exposure. The radiotracer dose is essential to PET image quality and quantification. Efforts are made to ensure the exact dose administered is used to create the patient image; but, no measures ensure the administered dose completely enters the patient circulation. Our objective was to use new technology (Lara, Lucerno Dynamics) to assess our infiltration rate, determine potential causes, reduce our rate, and assess results sustainability. Methods: Our IRB determined the project did not meet the definition of research as defined by 45 CFR 46.102(d) and classified the initiative as “quality improvement”. In Phase 1, our PET/CT center monitored the injection process. Lara data were analyzed (SAS Enterprise Miner, v. 14.1 and v.9.4) and identified potential contributing factors. We then implemented a quality improvement plan to address these factors. In Phase 2 we remeasured our infiltration rate. After Phase 2, we assessed our rate over an extended period. Results: Lara added 30 seconds to patient imaging. In Phase 1, 263 injections were monitored, and 35 infiltrations were identified (13.3%). Five technologists’ rates ranged from 8%-19%. Patients less than 145lbs with non-antecubital fossa injection sites were highly associated with infiltrations. In Phase 2, 278 injections were monitored, and 8 infiltrations were identified (2.9%), a 78% decrease (p < 0.0001). The same technologists’ rates ranged from 0%-4%. Eight technologists performed the next 504 injections and six were infiltrated (1.2%). Conclusions: Our findings supported published PET/CT infiltration rates. Using new technology provided technologists with feedback on injections and lead to significant and sustainable improvements quickly, with minimal patient/procedure disruption. Infiltrations are a quality and safety issue. As a result, the injection process requires ongoing monitoring.

Classification of nicotine-dependent users in India: a decision-tree approach

India has been predicted to have the fastest increase in deaths attributable to tobacco in the first 2 decades of the twenty-first century. Consequently, it is imperative to examine the extent of nicotine dependency among adults in India. The main objective of the present article is to characterize nicotine dependency related to smoking and smokeless tobacco among adults according to age, education level, duration of use and other socioeconomic characteristics. We analyzed Global Adult Tobacco Survey (GATS) 2010 data for 9190 smokers and 13,357 smokeless tobacco users who were 15 years of age or older. Time to first tobacco use of the day was used as a measure of nicotine dependence. We employed a decision tree algorithm from SAS Enterprise Miner to conduct classification analysis to establish the relationship among nicotine dependency for smoking, smokeless tobacco use, and its predictors. More than 65% of smokers and 54.3% of smokeless tobacco users were nicotine dependent. The decision tree results showed that the most important explanatory variable for the prediction of smoking dependency was duration of smoking followed by education, gender and region. For smokeless tobacco use, duration of smokeless tobacco use was also the most important explanatory variable followed by education, occupation and age. Smoking and smokeless tobacco use dependencies are driven by different sets of demographic and socioeconomic predictors. Suitable plans for tobacco cessation need to be implemented based on the type of tobacco and particularly duration of use, which was found to be the most important determinant of quitting.

Predictive Modeling for Imbalanced Big Data in SAS Enterprise Miner and R

There have been a variety of predictive models capable of handling binary targets, ranging from traditional logistic regression to modern neural networks. However, when the target variable represents a rare event, these models might not be appropriate as they assume that the distribution in the target variable is balanced. In this article, the impact of multiple resampling methods on conventional predictive models is studied. These resampling techniques include the methods of oversampling of the rare events, undersampling of the common events in the data, and synthetic minority over-sampling technique (SMOTE). The predictive models of decision trees, logistic regression and rule induction are applied with SAS Enterprise Miner (EM) software to the revised data. The studied data set is of home mortgage applications which includes a target variable with an occurrence rate of the rare event being 0.8%. The authors varied the percentage of the rare event from the original of 0.8% up to 50% and monitored the associated performances of the three predictive models to see which one worked the best.

Beyond Words of Popularization Mining Reviews on Comic Books Movies

Objectives: This paper aims to analyse the terms used by the of movie critics on popular comic book movies. From the analysis we are able to know what are the factors that most movie critics will focus at, whether the top movie critics review has an impact on the decision of general movie goers and from this will determine why the movies was a box office success. Methods: The research is done via text mining from SAS Enterprise Miner software which utilize the text analytics tool to analyse movie reviews by movie critics’ professionals. The analytical processes involved are text parsing, text filtering, text clustering and the text topic method. Correlation of terms is identified to help in determining the significance of each terms nested within the reviews. Findings: The findings show that the actors, the movie's characters and the film storyline are the most common type of terms that most movie critics are focusing on, thus this factors is what drives the reputation of the film and making it a blockbuster hit around the world. Application/Improvements: The results obtained can help the entertainment industry on their decision making on what to focus on when it comes to producing comic books superheroes into movies based on the sentiment analysis in identifying the postive and negative terms and their relationship with one another.

Predictive Modeling of Nonlinear Non-Stationary Processes in Crop Production Using Tools of SAS Enterprise Miner

Blackground. The issue of providing the increase of production of main agricultural crops inUkraine under conditions of environmental management requires the use of modern scientific approaches. The complexity of solving this problem lies in the lack of practical experience of applying modern information-analytical systems, where different methods for analysis and modeling of nonlinear non-stationary processes in crop production would be implemented simultaneously. The proposed methodology has the advantage of using the tools of SAS Enterprise Miner – software where a wide range of techniques are implemented, that should be used for predictive modeling of main agricultural crops according to the performed research.Objective. The goal of the study is in application of the integrated methods of analysis and predictive modeling of non-stationary processes for agricultural crop yield prediction using SAS Enterprise Miner tools.Methods. To solve the problems stated the following approaches were used: systems analysis, regression analysis, gradient boosting, probabilistic modeling and decision trees. The methodology for developing of crop yield prediction under influence of various groups of factors was offered, and the possibility of their use in decision support systems in agriculture was substantiated.Results. Based on the analysis of the works of domestic and foreign scientists it was proposed to improve methodology of development of yield prediction of main agricultural crops using integrated analysis methods, which were implemented in the system of SAS Enterprise Miner. The analysis of the obtained results was performed.Conclusions. Winter wheat and corn yield prediction was performed for the Forest-Steppe Zone using the developed methodic. Different methods of construction of models for prediction of the non-stationary processes were applied; the choice of the worthiest one was reasonably proved. Advanced information technologies, including SAS Enterprise Miner, were used for automatization the process of selecting the optimal model for investigated crop yield prediction.

Enabling self-service BI: A methodology and a case study for a model management warehouse

The promise of Self-Service Business Intelligence (BI) is its ability to give business users access to selection, analysis, and reporting tools without requiring intervention from IT. This is essential if BI is to maximize its contribution by radically transforming how people make decisions. However, while some progress has been made through tools such as SAS Enterprise Miner, IBM SPSS Modeler, and RapidMiner, analytical modeling remains firmly in the domain of IT departments and data scientists. The development of tools that mitigate the need for modeling expertise remains the “missing link” in self-service BI, but prior attempts at developing modeling languages for non-technical audiences have not been widely implemented. By introducing a structured methodology for model formulation specifically designed for practitioners, this paper fills the unmet need to bring model-building to a mainstream business audience. The paper also shows how to build a dimensional Model Management Warehouse that supports the proposed methodology, and demonstrates the viability of this approach by applying it to a problem faced by the Division of Fiscal and Actuarial Services of the US Department of Labor. The paper concludes by outlining several areas for future research.