作为类似于经典参数似然的非参数方法，经验似然以及由其拓展的非参似然方法是近年来统计学的一个重要研究领域和热点问题；此外在实际中，舍入、高维和相依等复杂数据的出现，对统计理论研究和应用提出了许多新的问题和挑战。.本项目拟借助经验似然及其推广、经验过程、投影追踪、Bootstrap、蒙特卡罗模拟、垂直密度表示和相依数据中的较新技术或工具，讨论经验似然方法在广义非参似然和Lq似然两个方面的推广、相应的统计性质和应用。项目首先构造推广的广义非参似然和非参Lq似然(均含经验似然)，然后在此基础上，讨论相应的估计和检验问题，进而把这些方法和工具应用到舍入、高维和相依等复杂数据方面。希望通过这些研究为实际数据处理工作者提供较优且切实可行的统计应用工具。

As a nonparametric analogue of the classical likelihood, empirical likelihood and its generalization has been one of the important research fields and hot disussions in statistics. At the same time, the complex data such as rounded data, high-dimensional data, and dependent data are frequently encountered in practice, and they bring many new questions and challenges in statistical theory studies and statistical applications.. By using the newest techniques in the fields as follows: empirical likelihood and its generalizations、empirical processes、projection pursuit、bootstrap、Monte Carlo simulation、vertical density representation and dependent data, we generalize the empirical likelihood and Lq likelihood to nonparametric statistics and investigate their statistical properties and applications. The project firstly constructs the generalized nonparametric likelihood and nonparametric Lq likelihood, then studies the estimate and test problems. Finally the methods are applied into the complex data such as rounded data, high-dimensional data, and dependent data. We hope these discussions can give more powerful and convenient techniques for the real data.