Research And Implementation Of Flight Dynamic Data Mining And Flight Prediction Model

With the advent of the cloud era, big data has attracted more and more attention. Internet of things, cloud computing, mobile internet, vehicle networking, phone, tablet, PC and a variety of sensors that throughout every corner of the earth, all of them are the sources of data or the ways of data carrier. Meanwhile, the core value of big data is the storage and analysis of massive data.With the rapid development of science and technology, the rapid increasement in flights and flights routes, the growing of air traffic flow, the result is a lot of flights history operating data. How to find some effective methods of data processing to effectively extracts and analyzes a wealth of knowledge and association rules that hidden in these data, it is a puzzle that has long plagued scientists. The emerging technology of massive data processing—Data Mining, provides a powerful tool for solving these problems.Firstly, this paper uses the way of data mining and the method of principal component analysis in feature extraction technique, selects/extracts/transformates/preprocesses/analysises and assimilates the original flight data, digs out the informations of busyness degree, the average time of flights, the delay time, the departure punctuality rate and the arrival punctuality rate. What’s more, according to the space tables published on the airlines’ websites, calculates the relative average prices of the space tables, estimates the total seats according to the maximum number of actual monthly airline passengers, calculates the attendance rates according to the number of passengers of China Southern Airlines 2013, then analyzes the correlation coefficient between the attendance rates and relative average prices/seasons.Secondly, this paper uses the prediction method of time series, establishes the model of linear regression/nonlinear regression/artificial neural network/bayesian network constant mean value. Based on prior information/the overall distribution of information/sample information, calculates the pattern of posterior distribution information. According to the current flight statuses and the history data of average flight time, adds experience and judgment of decision makers, estimates the possible arrival time of the flights.At last, through the analysis of experimental results, we can confirm that, using the method of data mining to add up the correlation coefficient between the attendance rates and the relative average prices/seasons; at the same time, using the prediction method of time series to estimate the possible arrival time of the flights, it has definite reference value in drawing up the flight planning time and the prices of the tickets.