军队离退休干部为党、国家和军队建设作出了重要贡献，妥善安置离退休干部是军地各级的重要政治任务。离退休干部移交政府安置（简称离退休干部安置），指军队和地方政府为离退休干部落实政治、生活等各项待遇的过程[1]。目前安置期为3年，即退休第1年审定安置去向（安置地），第2年落实安置住房，第3年办理安置手续。某军区近30多年来共安置1万余名离退休干部，但现行安置模式下安置去向确定时间晚，导致安置期工作忙乱，落实质量不高，造成部分离退休干部滞留部队，还引发一些矛盾纠纷。因此，将数据挖掘技术引入离退休干部安置工作，通过对离退休干部安置数据分析研究，发掘安置地确定规律、构建安置地预测模型，对于推进安置工作良性运行、促进军队干部队伍科学发展具有重要现实意义。论文主要工作如下： 首先，对离退休干部数据进行预处理，通过数据清理、数据变换、数据规约、属性构造，形成离退休干部安置数据集，并运用SPSS Statistics平台进行基础分析，得到了数据分布和安置地确定基本规律。 其次，对离退休干部安置数据集进行关联规则挖掘，在SPSS Clementine平台中运用Apriori算法、GRI算法获取关联规则，并从实践层面予以描述和解释。在此基础上，采取加入辅助变量的方法改进了挖掘模式。 最后，构建离退休干部安置数据预测模型，在SPSS Clementine平台中运用决策树、回归分析、神经网络建模，并进行分析、测试和评估。在此基础上优选预测模型，并采取剔除特殊数据集的办法予以改进，形成了较高预测准确率的模型。 本文研究成果运用于安置工作实践，可以在干部退休前就较为准确地预测安置地，从而前移安置工作起点，为安置期争取更多的时间，提高安置工作质量效益。此外，研究成果可以为正在进行的军队干部政策制度改革提供决策依据。

The retired military officers have made outstanding contributes during the construction of party, state and army. How to arrange those retired officers properly is a significant political task for armies and local governments at different levels. The transfern and resettlement of the retired military officers to the government, called the resettlement of the retired military officers for short, is to implement various treatments, like politics and livelihood, for the retired officers with the help of the military and local government. According to the current policy, a three-year period is required to complete the resettlement task. Specifically, the resettlement place and the housing arrangement are determined and implemented, respectively, in the first and the second years after retirement, and all resettlement procedures are finished in the last year. More than ten thousand retired militaries in a region have been resettled in the past three decades. However, under the current resettlement mode, the delayed determination of resettlement places leads to a busy period of resettlement and increases the difficulty of implementation, which in turn results in the backlog of retired military officers and even causes contradictions and disputes. Therefore, it has important realistic significance for carrying out the resettlement work effectively and constructing the army scientifically to understand the law of the determination of the resettlement places and to construct the predict model of resettlement places. This thesis introduces the data mining technique into the research of retired military officer resettlement for solving those problems. Firstly, the sets of data on the retired officers are obtained after the pre-processing of data, including data cleansing, data transformation, data specification, and property construction. Then, the law of data distribution and resettlement place determination is constructed by carrying out the basic analysis via the SPSS Statistics platform. Secondly, the association rules mining of the sets of data on the retired officers is carried out. The association rules are obtained by using the Apriori algorithm and the GRI algorithm in the SPSS Clementine platform and they are descripted and explained from the practical level. In addition, the improved mining scheme is proposed by introducing the auxiliary variables. Finally, the predict model of resettlement data for the retired officers is constructed. The analysis, test, and assessment for different predict models are carried out by using the decision tree, regression analysis, and neural network modelling in the SPSS Clementine platform. Then, by choosing the model with the best performance and further improving it via deleting the unusual data set, the predict model with high predict accuracy is constructed. By using the research achievement of this thesis during the practical resettlement of retired military officers, the resettlement places of upcoming retired officers can be accurately predicted before their retirements, and the resettlement work can be started early for a longer-period of resettlement, which further promotes the resettlement more effective and increases the quantity and quality of resettlement. Moreover, the research achievement can also provide a decision guideline for the ongoing reform of the military cadre policies.