The usual convention for assigning a treatment to an individual is a "one-size fits all" rule that is based on broad spectrum trends. Heterogeneity within and between subjects and improvements in scientific research convey the need for more effective treatment assignment strategies. Precisionized treatment (PT) offers an alternative to the traditional treatment assignment approach by making treatment decisions based on one or more covariates pertaining to an individual. We investigate two methods to inform PT rules: the Maximum Likelihood Estimation (MLE) method and the Boosting method. We apply these methods in the context of a crossover study design with a continuous outcome variable, one continuous covariate, and two intervention options. We explore the methods via extensive simulation studies and apply them to a data set from a study of safety warnings in passenger vehicles. We evaluate the performance of the estimated PT rules based on the improvement in mean response (RMD), the percent of correct treatment assignments (PCC), and the accuracy of estimating the location of the crossing point (MSE((x_c )). We also define a new metric that we call the percent of anomalies (PA). We characterize the potential benefit of using PT by relating it to the strength of interaction, the location of the crossing point, and the within-person intraclass correlation (ICC). We also explore the effects of sample size and overall variance along with the methods’ robustness to violations of model assumptions. We investigate the performance of the Boosting method under the standard weight and two alternative weighting schemes. Our investigation indicated the largest potential benefit of implementing a PT approach was when the crossover point was near the median, the strength of interaction was large, and the ICC was high. When a PT rule is used to assign treatments instead of a one-size fits all rule, an approximate 10-30% improvement in mean outcome can be gained. The MLE and Boosting method performed comparably across most of the simulation scenarios, yet in our data example, it appeared there may be an empirical benefit of the Boosting method over the MLE method. Under a distribution misspecification, the difference in performance between the methods was minor; however, when the functional form of the model was misspecified, we began to see improvement of the Boosting method over the MLE method. In the simulation conditions we considered, the weighting scheme used in the Boosting method did not markedly impact performance. Using data to develop PT rules can lead to an improvement in outcome over the standard approach of assigning treatments. We found that in a variety of scenarios, there was little added benefit to utilizing the more complex iterative Boosting procedure compared to the relatively straightforward MLE method when developing the PT rules. The results from our investigations could be used to optimize treatment recommendations for participants in future studies.

Orientador: Carlos Roberto Padovani / Banca: Adriano Wagner Ballarin / Banca: Flávio Fekkari Aragon / Banca: José Carlos Martinez / Banca: Marie Oshiiwa / Resumo: Aspectos teóricos relacionados à Análise Discriminante Multivariada - Linear e Quadrática - foram discutidos, por meio de um extenso levantamento histórico da função discriminante, com seus primórdios no trabalho de Fisher e sua posterior evolução, enfocando o intenso desenvolvimento das técnicas classificatórias discriminantes com o advento dos computadores. Foi dada ênfase aos softwares estatísticos desenvolvidos para PC, que realizam a análise discriminante, e que representam uma grande contribuição para pesquisadores e usuários desta técnica. Considerando a dificuldade existente quanto a aplicativos computacionais acessíveis a pesquisadores da área de ciências agrárias, elaborou-se um programa que realiza a análise discriminante quadrática com as respectivas freqüências de classificação correta, bem como o manual explicativo do usuário. Verificou-se que a função discriminante quadrática trata de um procedimento bastante útil nas ciências agrárias, como, por exemplo, em estudos nas áreas de solos, cultivos diversos (soja, milho, cana de açúcar, pupunha, braquiária, frutas), criação de animais e classificação e seleção de madeiras; porém, subutilizada frente à dificuldade de programas computacionais de fácil manuseio e acesso a pesquisadores das áreas aplicadas. Os procedimentos estudados e discutidos foram ilustrados com exemplos de aplicação, utilizando dados experimentais agronômicos de espécies de Girassóis e Eucalyptus, submetidos ao aplicativo desenvolvido. / Abstract: A large historical study of the discriminant function has allowed a discussion on theoretical aspects related to the Multivaried Discriminant Analysis - Linear and Quadratic, showing its past in the work of Fisher and its later evolution, emphasizing the wide development of classificatory discriminant techniques with the happening of the computers, and specific statistic softwares which practice the discriminant analysis, representing a big contribution to researches and users of this technique. Considering the difficulty in relation to accessible softwares to researches of the agrarian area, a software which performs a linear and quadratic discriminant analysis was built with its frequencies of correct classification, as well as an explicative manual to users. The quadratic discriminant was studied as being a very useful process in agrarian sciences. Some examples of this usefulness is in studies of the ground, diversified cultivation (soybean, corn, sugarcane, pejibaye, brachiaria decumbens fruits), animal creation and wood selection, and classification; however, misused in relation to the difficulties of easy handing and access to researchers of applied areas. The studied and discussed procedures were illustrated with applications, using agronomic experimental data of Sunflower and Eucalyptus, submitted to developed software. / Doutor

Made available in DSpace on 2014-06-11T19:31:35Z (GMT). No. of bitstreams: 0 Previous issue date: 2006-12-19Bitstream added on 2014-06-13T21:02:54Z : No. of bitstreams: 1 simeao_sfap_dr_botfca.pdf: 899191 bytes, checksum: da6ed77a45734c278c56395d23c51cd0 (MD5) / Universidade Estadual Paulista (UNESP) / Aspectos teóricos relacionados à Análise Discriminante Multivariada - Linear e Quadrática - foram discutidos, por meio de um extenso levantamento histórico da função discriminante, com seus primórdios no trabalho de Fisher e sua posterior evolução, enfocando o intenso desenvolvimento das técnicas classificatórias discriminantes com o advento dos computadores. Foi dada ênfase aos softwares estatísticos desenvolvidos para PC, que realizam a análise discriminante, e que representam uma grande contribuição para pesquisadores e usuários desta técnica. Considerando a dificuldade existente quanto a aplicativos computacionais acessíveis a pesquisadores da área de ciências agrárias, elaborou-se um programa que realiza a análise discriminante quadrática com as respectivas freqüências de classificação correta, bem como o manual explicativo do usuário. Verificou-se que a função discriminante quadrática trata de um procedimento bastante útil nas ciências agrárias, como, por exemplo, em estudos nas áreas de solos, cultivos diversos (soja, milho, cana de açúcar, pupunha, braquiária, frutas), criação de animais e classificação e seleção de madeiras; porém, subutilizada frente à dificuldade de programas computacionais de fácil manuseio e acesso a pesquisadores das áreas aplicadas. Os procedimentos estudados e discutidos foram ilustrados com exemplos de aplicação, utilizando dados experimentais agronômicos de espécies de Girassóis e Eucalyptus, submetidos ao aplicativo desenvolvido. / A large historical study of the discriminant function has allowed a discussion on theoretical aspects related to the Multivaried Discriminant Analysis - Linear and Quadratic, showing its past in the work of Fisher and its later evolution, emphasizing the wide development of classificatory discriminant techniques with the happening of the computers, and specific statistic softwares which practice the discriminant analysis, representing a big contribution to researches and users of this technique. Considering the difficulty in relation to accessible softwares to researches of the agrarian area, a software which performs a linear and quadratic discriminant analysis was built with its frequencies of correct classification, as well as an explicative manual to users. The quadratic discriminant was studied as being a very useful process in agrarian sciences. Some examples of this usefulness is in studies of the ground, diversified cultivation (soybean, corn, sugarcane, pejibaye, brachiaria decumbens fruits), animal creation and wood selection, and classification; however, misused in relation to the difficulties of easy handing and access to researchers of applied areas. The studied and discussed procedures were illustrated with applications, using agronomic experimental data of Sunflower and Eucalyptus, submitted to developed software.