Temperature Estimation with Time Series Analysis from Air Quality Data Set

With the expansion of the data size, data mining techniques are gaining more and more importance. Data mining consists of methods such as classification, clustering, time series estimation and association rule. In this study, a time series analysis is carried out in order to make an estimation for the future in accordance with the structure of the data set. Time series are series in which the variables are recorded in chronological order. The data set was created by recording the gas concentrations in the air at a time interval. These data are used to estimate the changes in air quality. Three types of time series analysis training algorithm are used in the study. The results given by the algorithms are close to each other and high performance has been determined. As a result of experimental studies, it is observed that time series analysis is sufficient to estimate air quality.

I. Introduction

Today, due to the expansion of the size of data sets, traditional methods are not useful for the analysis of this data. Data mining methods are the process of extracting meaningful information from these large data sets. The large data sets in many scientific areas can be used for the benefit of humanity through these methods. The methods used in data mining are classification, clustering and association rule. Recently, time series analysis in classification algorithms has become the most preferred analysis.