Submitted by Mario BC (mario@bc.ufrpe.br) on 2016-12-02T13:12:59Z No. of bitstreams: 1 Gabriela del Carmen Calderon Estrada.pdf: 1790318 bytes, checksum: 59a9ef3279b882660365d852f8a0f3a1 (MD5) / Made available in DSpace on 2016-12-02T13:12:59Z (GMT). No. of bitstreams: 1 Gabriela del Carmen Calderon Estrada.pdf: 1790318 bytes, checksum: 59a9ef3279b882660365d852f8a0f3a1 (MD5) Previous issue date: 2015-12-11 / Conselho Nacional de Pesquisa e Desenvolvimento Científico e Tecnológico - CNPq / The rust, caused by the fungus Hemileia vastatrix Berk & Br., is the main disease of coffee (Coffea arabica L.) in Latin America. The principal damage caused is defoliation and death of lateral branches, which causes premature fruit losses. Guatemala produces coffee in 270,000 hectares, and near of the 82% is cultivated with susceptible varieties to coffee rust races. Coffee rust epidemic is a complex process based on the relationships between the environment, plant growth, and crop practices. The objective of this study was to develop models for risk analysis based on decision trees in order to understand how cropping patterns determine the progress of the disease in Guatemala to identify and prioritize the important factors. For this work were used 1215 observations, obtained in 35 coffee plots from April 2013 to December 2014. The modeled variable was the leaf severity. Using the CHAID (Chi-Square Automatic Interaction Detection) algorithm were developed two decision trees. The first predicts leaf severity in plots where the producer does not follow the disease, while the second requires rust monitoring 28 days before the date of the severity risk analysis. In the trees, the main predictor was the fungicide spraying per year. The following predictor variables on the tree were related with the tissue availability for new infections, which also stimulates microenvironments with high relative humidity, warm temperatures, and foliar wetness prevalence. Only for non-monitoring tree was included the average rainfall, which suggests that climate relationship with the epidemic, is at microclimate level. The tree for plots with disease monitoring includes in all levels the 28 before severity and replaced management or climate variables getting similar predicted values. The accuracy of the tree for monitored plots was 65.85% with an estimated accuracy by cross validation of 73.34%, and for the monitored plots, the accuracy was 62.53% and 68.54%, respectively. Risk analysis models prove to be tools of support in making management decisions to implement the control of coffee rust and allow list in order of importance, management practices, and climatic factors that influence disease severity in different crop patterns. / A ferrugem do cafeeiro, causada pelo fungo Hemileia vastatrix Berk & Br., é a principal doença do cafeeiro (Coffea arabica L.) na América Latina. O principal dano é desfolha e morte de ramos laterais, que provocam perdas prematuras de frutos. A Guatemala produz café em 270.000 hectares, sendo que cerca de 82% é cultivado com variedades suscetíveis às raças de ferrugem. A epidemia da ferrugem é um processo complexo baseado nas relações entre ambiente, crescimento da planta, e práticas de manejo. O objetivo deste estudo foi desenvolver modelos para análise de risco baseados em árvores de decisão, a fim de entender como os padrões de cultivo determinam o progresso da doença na Guatemala para identificae e priorizar os fatores importantes. Para este trabalho foram utilizadas 1215 observações, obtidas de 35 lavouras de abril de 2013 a dezembro de 2014. A variável modelada foi a severidade da folha. Utilizando o algoritmo CHAID (Chi-Quadrado Detecção Automatic Interaction), foram desenvolvidas duas árvores de decisão. A primeira árvore permite prever a severidade na folha nas parcelas em que o produtor não realiza acompanhamento da doença, enquanto a segunda requer o monitoramento da ferrugem 28 dias antes da data da análise de risco da severidade. Nas árvores, o principal preditor foi o número de aplicações de fungicida por ano. As seguintes variáveis preditoras na árvore foram relacionadas com disponibilidade de tecido para novas infecções, que podem favorecem a formação de microambientes com alta umidade relativa, temperaturas amenas e prevalência da molhadura folhar. Apenas para a árvore de não monitoramento foi incluída a variável da precipitação média, o que sugere que a relação do clima é em nível microclimático. A árvore com monitoramento inclui em todos os níveis a severidade aos 28 dias antes e substitui variáveis de manejo ou clima, estimando valores semelhantes. A acurácia da árvore para lavouras não monitoradas foi de 65,85% com uma estimativa de acurácia por validação cruzada de 73,34%. Na árvore para lavouras monitoradas a acurácia foi de 62,53% e 68,54%, respectivamente. Os modelos de análise de risco demonstram ser ferramentas de apoio na tomada de decisões de manejo para implementar o controle da ferrugem do cafeeiro e possibilitam listar, em ordem de importância, as práticas de manejo e fatores climáticos que influenciam na severidade da doença em diferentes padrões do cultivo.

This study described self-injury within a general adolescent population. This study involved secondary analysis of data gathered using the middle school Youth Risk Behavior Survey (YRBS) from 1,748 sixth- and eighth-grade students in eight middle schools in a large, southeastern county in Florida. A substantial percentage of students surveyed (28.4%) had tried self-injury. The prevalence of having ever tried self-injury did not vary by race or ethnicity, grade, school attended, or age but did differ by gender. When controlling for all other variables in the multivariate model including suicide, having ever tried self-injury was associated with peer self-injury, inhalant use, belief in possibilities, abnormal eating behaviors, and suicide scale scores. Youth who knew a friend who had self-injured, had used inhalants, had higher levels of abnormal eating behaviors, and higher levels of suicidal tendencies were at increased risk for having tried self-injury. Youth who had high belief in their possibilities were at decreased risk for having tried self-injury. During the past month, most youth had never harmed themselves on purpose. Approximately 15% had harmed themselves one time. Smaller proportions of youth had harmed themselves more frequently, including two or three different times (5%), four or five different times (2%), and six or more different times (3%). The frequency of self-injury did not vary by gender, race or ethnicity, grade, or school attended. Almost half of students surveyed (46.8%) knew a friend who had harmed themselves on purpose. Peer self-injury demonstrated multivariate relationships with gender, having ever been cyberbullied, having ever tried self-injury, grade level, and substance use. Being female, having been cyberbullied, having tried self-injury, being in eighth grade, and higher levels of substance use placed youth at increased risk of knowing a peer who had self-injured. Chi-squared Automatic Interaction Detection (CHAID) was used to identify segments of youth at greatest and least risk of self-injury, frequent self-injury, and knowing a friend who had harmed themselves on purpose (i.e., peer self-injury).

The purpose of this dissertation is to understand why individuals enrolled in the Special Supplemental Nutrition Program for Women, Infants and Children (WIC) fail to retrieve food vouchers, miss WIC appointments, and become inactive in program components. In Kentucky, mothers who fail to pick up food instruments for 60 days are automatically terminated from the program. The specific research questions that guided this study are: (1) Which segments of enrollees are at greatest and least risks of nonparticipation in the WIC program? (2) How do predisposing, enabling and need characteristics impact WIC nonparticipation among eligible mothers? (3) How do WIC enrollees describe their experiences using WIC? (4) What do WIC enrollees report as reasons for nonparticipation while still eligible? Addressing these research questions will inform the development of practical outreach solutions specifically tailored for the purpose of mitigating nonparticipation in WIC and contribute to our understanding of the factors that deter eligible families from using government assistance programs like WIC.

The purpose of this study was to describe the predictors of evacuation intention among coastal residents in the State of Florida and to determine if there are meaningful segments of the population who intend to evacuate when told to do so by governmental officials because of a major hurricane. In the America’s and the Caribbean, 75,000 deaths have been attributed to hurricanes in the 20 th century. A well planned evacuation can reduce injury and death, yet many people do not have an evacuation plan and do not intend to evacuate when told to do so. The study used secondary data from the Harvard School of Public Health, Hurricane in High Risk Areas study, a random sample of 5,046 non-institutionalized persons age 18 and older in coastal counties of Texas, Louisiana, Mississippi, Alabama, Georgia, North Carolina, South Carolina and Florida. Surveys for the State of Florida were segregated and used in this analysis, resulting in a study sample of 1,006 surveys from 42 counties. When asked if they would evacuate in the future if told to by government officials, 59.1% of Floridians surveyed said they would leave, 35.2% said they would not leave and 5.6% said it would depend. In Florida, 65.7% of the population had been threatened or hit by a major hurricane in the last three years and 26.6% of those had left their homes because of the hurricane. Of those whose communities were threatened by a hurricane, 83.3% of the communities were damaged and 33.8% experienced major flooding associated with the hurricane. Bivariate statistics and logistic regression were used to explore the interactions of predictors and evacuation intention. The best predictor of evacuation intention was prior evacuation from a hurricane (chi-square= 45.48, p < .01, Cramer’s V = 0.266). Significant relationships were also demonstrated between evacuation intention and worry a future hurricane would hit the community (chi-square = 22.75, p < .01, Cramer’s V = 0.11), the presence of pets (chi-square = 6.57, p < .01, Cramer’s V = 0.084), concern the home would be damaged (chi-square = 19.41, p < .01, Cramer’s V = 0.10), belief the home would withstand a major hurricane (chi-square = 19.55, p < .01, Cramer’s V = 0.10), length of time in the community (chi-square = 26.59, p < .01, Cramer’s V = 0.12), having children in the household (chi-square = 11.13, p < .01, Cramer’s V = 0.11), having a generator (chi-square = 17.12, p < .01, Cramer’s V = 0.13), age (chi-square = 24, p < .01, Cramer’s V = 0.16) and race (chi-square = 12.21, p = .02, Cramer’s V = 0.12). Logistic regression of the predictors of evacuation intention resulted in significant relationships with previous evacuation experience (OR = 4.99, p < .001), age 30 to 49 compared to age over 65 (OR = 2.776, p < .01), the presence of a generator (OR = .447, p < .01), having a home not very likely to be damaged compared to a home very likely to be damaged (OR =.444, p = .018), and experiencing poor prior government and voluntary agency response to previous hurricanes compared to excellent response (OR = .386, p < .027). Chi-squared Automatic Interaction Detection (CHAID) was used to identify segments of the population most likely and least likely to evacuate when told to do so. Those most likely to evacuate had evacuated due to a previous hurricane. Those least likely to evacuate when told to do so had not evacuated in a previous storm, do not own a generator and are over the age of 65. Information from this study can be used in planning for evacuation response by governmental entities. Available demographic information can be used to determine numbers of persons likely to evacuate before a storm. The results of this study can be used to inform a marketing strategy by government officials to encourage evacuation among those who say they would not evacuate when told to do so. Further research is needed to determine additional characteristics of the populations who say they will and will not evacuate when told to do so.

Yes / Social media provides a significant avenue for stakeholder engagement which is crucial to ensure loyalty and satisfaction of stakeholders who possess valuable resources that can influence the business outcomes. Container lines – imperative members of global supply chains and facilitators of international trade – utilize social media to engage their stakeholders due to environmental and commercial complexity of their business. However, not all social media posts generate the same amount of stakeholder engagement. This study aims to identify and examine the social media post characteristics that lead to higher stakeholder engagement in the container shipping market. The study applies Chi-Squared Automatic Interaction Detection method to categorize social media posts based on their engagement levels. The analysis is conducted on the tweets of four global container lines which are posted between 1 September 2018 and 31 January 2019. The results demonstrate that social media posts of container lines have varying effects on engagement level. We found that fluency of tweets, tangibility of company resources in the tweet, vividness level, content type, existence of a link, and existence of a call-to-action significantly influence the container lines’ stakeholder engagement rate. This study is the first that finds out social media post classes based on the interaction between their characteristics and engagement rates by employing a decision tree methodology. The results are expected to help container lines in their social media management and stakeholder engagement policies.

O interesse pela área de análise e caracterização de materiais biomédicos cresce, devido a necessidade de selecionar de forma adequada, o material a ser utilizado. Dependendo das condições em que o material será submetido, a caracterização poderá abranger a avaliação de propriedades mecânicas, elétricas, bioatividade, imunogenicidade, eletrônicas, magnéticas, ópticas, químicas e térmicas. A literatura relata o emprego da técnica de árvores de decisão, utilizando os algoritmos SimpleCart(CART) e J48, para classificação de base de dados (dataset), gerada a partir de resultados de artigos científicos. Esse estudo foi realizado afim de identificar características superficiais que otimizassem a atividade celular. Para isso, avaliou-se, a partir de artigos publicados, o efeito de tratamento de superfície do titânio na atividade celular in vitro (células MC3TE-E1). Ficou constatado que, o emprego do algoritmo SimpleCart proporcionou uma melhor resposta em relação ao algoritmo J48. Nesse contexto, o presente trabalho tem como objetivo aplicar, para esse mesmo estudo, os algoritmos CHAID (Chi-square iteration automatic detection) e CHAID Exaustivo, comparando com os resultados obtidos com o emprego do algoritmo SimpleCart. A validação dos resultados, mostraram que o algoritmo CHAID Exaustivo obteve o melhor resultado em comparação ao algoritmo CHAID, obtendo uma estimativa de acerto de 75,9% contra 58,6% respectivamente, e um erro padrão de 7,9% contra 9,1% respectivamente, enquanto que, o algoritmo já testado na literatura SimpleCart(CART) teve como resultado 34,5% de estimativa de acerto com um erro padrão de 8,8%. Com relação aos tempos de execução apurados sobre 22 mil registros, evidenciaram que o algoritmo CHAID Exaustivo apresentou os melhores tempos, com ganho de 0,02 segundos sobre o algoritmo CHAID e 14,45 segundos sobre o algoritmo SimpleCart(CART). / The interest for the area of analysis and characterization of biomedical materials as the need for selecting the adequate material to be used increases. However, depending on the conditions to which materials are submitted, characterization may involve the evaluation of mechanical, electrical, optical, chemical and thermal properties besides bioactivity and immunogenicity. Literature review shows the application decision trees, using SimpleCart(CART) and J48 algorithms, to classify the dataset, which is generated from the results of scientific articles. Therefore the objective of this study was to identify surface characteristics that optimizes the cellular activity. Based on published articles, the effect of the surface treatment of titanium on the in vitro cells (MC3TE-E1 cells) was evaluated. It was found that applying SimpleCart algorithm gives better results than the J48. In this sense, the present study has the objective to apply the CHAID (Chi-square iteration automatic detection) algorithm and Exhaustive CHAID to the surveyed data, and compare the results obtained with the application of SimpleCart algorithm. The validation of the results showed that the Exhaustive CHAID obtained better results comparing to CHAID algorithm, obtaining 75.9 % of accurate estimation against 58.5%, respectively, while the standard error was 7.9% against 9.1%, respectively. Comparing the obtained results with SimpleCart(CART) results which had already been tested and presented in the literature, the results for accurate estimation was 34.5% and the standard error 8.8%. In relation to execution time found through the 22.000 registers, it showed that the algorithm Exhaustive CHAID presented the best times, with a gain of 0.02 seconds over the CHAID algorithm and 14.45 seconds over the SimpleCart(CART) algorithm.

隨著金融自由化、國際化的趨勢，加上國民所得提高、電子化的普及，使得信用卡市場蓬勃發展，國內各大銀行紛紛積極投入信用卡發卡行列。台灣的信用卡市場競爭的程度，從各發卡銀行所提供消費者的各項附加服務，如辦卡送禮、持卡免年費、失卡零風險、購物優惠…等，幾乎都已是每一張信用卡的基本配備。 隨著卡債、卡奴的事件爆發，銀行業者舊有的信用卡行銷策略已經宣告失敗，但信用卡市場背後帶來的經濟效益，仍然是不容忽視，如今，要如何增加信用卡市場的佔有率已不是銀行業者的行銷重點，高佔有率並不一定就能帶來高經濟效益。銀行業者的行銷策略應該是做好信用卡市場區隔，找出不同特性的消費族群，依消費族者選擇信用卡的考量因素擬定行銷策略，進而提升市場競爭地位。 本研究選用四種模型建置方式，分別為羅吉斯迴歸、C5.0、CHAID以及類神經網路，經由分類矩陣評估比較四種模型，其中C5.0不論是在整體預測正確率、反查率或準確度，皆是高於其它三個模型，故最後選擇C5.0此一模型。 透過C5.0共獲得七項影響「是否有使用信用卡」之相關變數，其中「是否有出國旅行」、「經濟來源是否為自己」、「性別」、「是否畢業後找工作」、「是否有使用網路消費」、「認同環保意識」、「是否有投資或買保險」，此七項變數對使用信用卡消費具較大影響力，最後本研究會針對這些變數再給與發卡銀行建議。 【關鍵字】信用卡、資料採礦、C5.0、CHAID、類神經網路 / As the trend of financial liberalization and globalization and also the popularization of electronic business and the increase of domestic income, the credit card market has bloomed vigorously then ever, banks are urging on developing credit card markets. All those additional service of every bank could be seen as a clue to know the competitiveness in Taiwan, such as free gift, free annual fee, zero risk of losing cards, shopping discount…etc., and those service almost become a basic equipment of every credit card. With credit debt and credit card slaves increasing, bank’s former marketing strategies have failed. The economic benefits of credit card market still are not ignored. Today, how to increase market share of credit card is not the key point of bank’s marketing strategy. There is not necessary that high market share can bring high economic benefits. In order to follow this trend, the study aims to discover the corn factors of possessing credit cards through the application of Clementine 12.0 software. Since Decision Tree-C5.0 is excellent in the forecast accuracy and validity as compared to Logistic Regression, Decision Tree-CHAID and Neural Net were adopted in this research. Through using Decision Tree-C5.0, this study identified seven factors that have greater impact on using credit cards and they are”Whether respondent travel abroad”,“Is the source of income making by yourself”,“Gender”,“Do respondent look for jobs after graduating from school”,“Do respondent buy something on the internet”,“Approve the environmental awareness”.This research will chiefly use these seven factors to provide the marketing portfolio strategy recommendations for banks. Keywords：Credit Card, Data Mining, C5.0, CHAID, Neural Net

O interesse pela área de análise e caracterização de materiais biomédicos cresce, devido a necessidade de selecionar de forma adequada, o material a ser utilizado. Dependendo das condições em que o material será submetido, a caracterização poderá abranger a avaliação de propriedades mecânicas, elétricas, bioatividade, imunogenicidade, eletrônicas, magnéticas, ópticas, químicas e térmicas. A literatura relata o emprego da técnica de árvores de decisão, utilizando os algoritmos SimpleCart(CART) e J48, para classificação de base de dados (dataset), gerada a partir de resultados de artigos científicos. Esse estudo foi realizado afim de identificar características superficiais que otimizassem a atividade celular. Para isso, avaliou-se, a partir de artigos publicados, o efeito de tratamento de superfície do titânio na atividade celular in vitro (células MC3TE-E1). Ficou constatado que, o emprego do algoritmo SimpleCart proporcionou uma melhor resposta em relação ao algoritmo J48. Nesse contexto, o presente trabalho tem como objetivo aplicar, para esse mesmo estudo, os algoritmos CHAID (Chi-square iteration automatic detection) e CHAID Exaustivo, comparando com os resultados obtidos com o emprego do algoritmo SimpleCart. A validação dos resultados, mostraram que o algoritmo CHAID Exaustivo obteve o melhor resultado em comparação ao algoritmo CHAID, obtendo uma estimativa de acerto de 75,9% contra 58,6% respectivamente, e um erro padrão de 7,9% contra 9,1% respectivamente, enquanto que, o algoritmo já testado na literatura SimpleCart(CART) teve como resultado 34,5% de estimativa de acerto com um erro padrão de 8,8%. Com relação aos tempos de execução apurados sobre 22 mil registros, evidenciaram que o algoritmo CHAID Exaustivo apresentou os melhores tempos, com ganho de 0,02 segundos sobre o algoritmo CHAID e 14,45 segundos sobre o algoritmo SimpleCart(CART). / The interest for the area of analysis and characterization of biomedical materials as the need for selecting the adequate material to be used increases. However, depending on the conditions to which materials are submitted, characterization may involve the evaluation of mechanical, electrical, optical, chemical and thermal properties besides bioactivity and immunogenicity. Literature review shows the application decision trees, using SimpleCart(CART) and J48 algorithms, to classify the dataset, which is generated from the results of scientific articles. Therefore the objective of this study was to identify surface characteristics that optimizes the cellular activity. Based on published articles, the effect of the surface treatment of titanium on the in vitro cells (MC3TE-E1 cells) was evaluated. It was found that applying SimpleCart algorithm gives better results than the J48. In this sense, the present study has the objective to apply the CHAID (Chi-square iteration automatic detection) algorithm and Exhaustive CHAID to the surveyed data, and compare the results obtained with the application of SimpleCart algorithm. The validation of the results showed that the Exhaustive CHAID obtained better results comparing to CHAID algorithm, obtaining 75.9 % of accurate estimation against 58.5%, respectively, while the standard error was 7.9% against 9.1%, respectively. Comparing the obtained results with SimpleCart(CART) results which had already been tested and presented in the literature, the results for accurate estimation was 34.5% and the standard error 8.8%. In relation to execution time found through the 22.000 registers, it showed that the algorithm Exhaustive CHAID presented the best times, with a gain of 0.02 seconds over the CHAID algorithm and 14.45 seconds over the SimpleCart(CART) algorithm.

O interesse pela área de análise e caracterização de materiais biomédicos cresce, devido a necessidade de selecionar de forma adequada, o material a ser utilizado. Dependendo das condições em que o material será submetido, a caracterização poderá abranger a avaliação de propriedades mecânicas, elétricas, bioatividade, imunogenicidade, eletrônicas, magnéticas, ópticas, químicas e térmicas. A literatura relata o emprego da técnica de árvores de decisão, utilizando os algoritmos SimpleCart(CART) e J48, para classificação de base de dados (dataset), gerada a partir de resultados de artigos científicos. Esse estudo foi realizado afim de identificar características superficiais que otimizassem a atividade celular. Para isso, avaliou-se, a partir de artigos publicados, o efeito de tratamento de superfície do titânio na atividade celular in vitro (células MC3TE-E1). Ficou constatado que, o emprego do algoritmo SimpleCart proporcionou uma melhor resposta em relação ao algoritmo J48. Nesse contexto, o presente trabalho tem como objetivo aplicar, para esse mesmo estudo, os algoritmos CHAID (Chi-square iteration automatic detection) e CHAID Exaustivo, comparando com os resultados obtidos com o emprego do algoritmo SimpleCart. A validação dos resultados, mostraram que o algoritmo CHAID Exaustivo obteve o melhor resultado em comparação ao algoritmo CHAID, obtendo uma estimativa de acerto de 75,9% contra 58,6% respectivamente, e um erro padrão de 7,9% contra 9,1% respectivamente, enquanto que, o algoritmo já testado na literatura SimpleCart(CART) teve como resultado 34,5% de estimativa de acerto com um erro padrão de 8,8%. Com relação aos tempos de execução apurados sobre 22 mil registros, evidenciaram que o algoritmo CHAID Exaustivo apresentou os melhores tempos, com ganho de 0,02 segundos sobre o algoritmo CHAID e 14,45 segundos sobre o algoritmo SimpleCart(CART). / The interest for the area of analysis and characterization of biomedical materials as the need for selecting the adequate material to be used increases. However, depending on the conditions to which materials are submitted, characterization may involve the evaluation of mechanical, electrical, optical, chemical and thermal properties besides bioactivity and immunogenicity. Literature review shows the application decision trees, using SimpleCart(CART) and J48 algorithms, to classify the dataset, which is generated from the results of scientific articles. Therefore the objective of this study was to identify surface characteristics that optimizes the cellular activity. Based on published articles, the effect of the surface treatment of titanium on the in vitro cells (MC3TE-E1 cells) was evaluated. It was found that applying SimpleCart algorithm gives better results than the J48. In this sense, the present study has the objective to apply the CHAID (Chi-square iteration automatic detection) algorithm and Exhaustive CHAID to the surveyed data, and compare the results obtained with the application of SimpleCart algorithm. The validation of the results showed that the Exhaustive CHAID obtained better results comparing to CHAID algorithm, obtaining 75.9 % of accurate estimation against 58.5%, respectively, while the standard error was 7.9% against 9.1%, respectively. Comparing the obtained results with SimpleCart(CART) results which had already been tested and presented in the literature, the results for accurate estimation was 34.5% and the standard error 8.8%. In relation to execution time found through the 22.000 registers, it showed that the algorithm Exhaustive CHAID presented the best times, with a gain of 0.02 seconds over the CHAID algorithm and 14.45 seconds over the SimpleCart(CART) algorithm.

Background and Significance: Cesarean delivery rates increased by more than 50% between 1996 and 2011 in the United States. The large increase in rates for the procedure was generally not associated with significant improvements in obstetric outcomes, raising concern about quality and prompting recommendations for prevention. Primary cesareans provide the best opportunity to reduce overall cesarean rates, and the group of first-time mothers considered low-risk for cesarean (known as nulliparous, term, singleton, vertex, NTSV) constitutes the focus of prevention efforts. Studies increasingly report racial and ethnic differences in NTSV cesareans, which remain after controlling for health factors. However, the reasons for these disparities and whether or not they can be mitigated are issues that are not well known. The objective of this investigation was to examine factors that modify the association between race, ethnicity and NTSV cesarean deliveries in Florida. Our overall aim was to improve understanding of drivers of racial and ethnic disparities in cesareans in order to inform efforts to reduce disparities. Methods: We conducted a population-based retrospective cohort study of 145,117 NTSV deliveries in labor, using Florida’s linked birth certificate and maternal hospital discharge records for the period of 2012 to 2014. The study was restricted to births in routine delivery hospitals to five racial and ethnic groups: non-Hispanic whites and blacks (including Haitians), Cubans, Puerto Ricans, and Mexicans. Two contrasting approaches were employed in the analysis. First, generalized linear mixed modelling was used to examine, quantify and describe effect modification of the race/ethnicity–association by cesarean risk factors. Non-Hispanic whites were the reference group for comparison. Second, classification tree modeling (chi-Squared Automatic Interaction Detection, CHAID) was used to identify cesarean risk factor combinations that define distinct subgroups with high and low rates of NTSV cesarean among the different racial and ethnic groups in the study population. Risk factors examined included individual socioeconomic, medical and health service-related factors, hospital factors, and a maternal neighborhood index of deprivation/affluence. Results: Non-Hispanic whites were the largest racial/ethnic group in the study population (57.6%), followed by non-Hispanic blacks (23%), Cubans (8.1%), Puerto Ricans (6.8%) and finally Mexicans (4.5%). All four minority groups experienced a higher risk of cesarean relative to non-Hispanic whites after adjusting for significant risk factors, with Cubans having the highest adjusted risk ratio (RR, 1.27) followed by non-Hispanic blacks (RR, 1.18). From the regression-based tests of effect modification, we found positive interactions between race (non-Hispanic black versus white), older gestational age, and labor induction; and negative interactions between ethnicity (Cuban versus non-Hispanic white), presence of medical risk conditions, and labor induction. The adjusted RR of cesarean comparing blacks to whites was 1.04 among spontaneous deliveries at early term (P=.33), but increased to 1.28 (P Conclusions: Our findings on risk factors that modified the association between race, ethnicity and NTSV cesarean delivery and differences in cesarean risk subgroups between racial and ethnic groups suggest that there are potential opportunities to reduce disparities in rates for the procedure in Florida. Whereas racial disparities appear to be related to disparities in health service factors related to cesarean, ethnic disparities appear to persist above and beyond the medical and health service factors included in this investigation. Further research, potentially involving qualitative methods and targeting some of the identified maternal subgroups with high rates of cesarean may help clarify maternal cultural factors, or differences in patient-provider interaction, that may contribute to some of the disparities.