Cash income from Land and/or Building Tax is the most important part for the development of a region, with the results obtained by the regional government being able to improve various infrastructures that help the community in carrying out various activities and make the area more developed . High compliance from taxpayers is a condition where taxpayers fulfill all their obligations in paying taxes, but in reality there are still challenges faced by taxpayers in reality, not all of them even have a low level of obedient taxpayers in paying taxes. Therefore, to optimize the management of land and building taxes, a good planning, organization, mobilization and supervision is needed by the Bandar Lampung City Government, especially for the Badan Pengelola Pajak dan Retribusi Daerah (BPPRD) agency in terms of collecting Land and Building Taxes.Methods used for classification is the Naive Bayes Algorithm where this algorithm is proven to have high accuracy and speed when applied to databases with large data. Nave Bayes algorithm is an algorithm that has a function to predict or predict an event in the future using previous data. Primary data from BPPRD Agency will then be validated using the RapidMiner tool.The results of research using data mining, grouping data or what is often referred to as data classification, is very useful for predicting which taxpayers will be smooth or not smooth in paying PBB taxes in Tanjung senang District. The calculation is done by determining the probability value of each attribute and then it will be validated with Ten-fold Cross Validation to determine the level of accuracy, precision, and recall of the classification prediction results.

Keywords: Tax, Cross Validation, Algoritma naïve bayes, RapidMiner.