چكيده به لاتين
Abstract
Nowadays, given the increasing spread of information that human beings are dealing with, it is necessary to use methods such as data mining to extract information contained in the data. Data mining is a kind of knowledge in which it can be interpreted by extraction of knowledge through the collection and preprocessing of data. In this thesis, the researcher collected data from 1224 people who referred to one of the Sari health centers and analyzed them using data mining algorithms to discover hidden patterns in diabetes data. One of the most important functions in data mining is health storage and mechanization of data. In this research, tests related to diabetes have been collected manually. These factors include height, weight, body mass index, and blood pressure and etc. Then, using information about diabetic patients, the data were normalized in the first step and in the next step, the missing data were identified using the Crisp preprocessing methodology and the incomplete data were recovered or deleted. In step 3, using a variety of data mining methods the effect of specificity factors on diabetes has been studied and the number of factors influencing diabetes has been identified. Finally, using some data mining algorithms (including decision tree, neural network, etc.), predicted the condition of diabetes. In this research, important factors affecting diabetes have been identified, including age, height, and weight، and the tree algorithm is the best algorithm to predict the diabetes disease
Keywords: Diabetes, Data mining, Neural network, Decision tree
