Please study condition forecasting of mobile data traffic from where? The teacher asked me to study, but did not give me the data, lost

CodePudding user response:

Home to predict more easily, but shall not apply to the foreign
I tell what, brother, you this 20 points is a bit of my typing time.
1, you are using ordinary programming language, or a spark? If large amount of data with the spark may than regular development language that
2, since be predict, so there must be historical data, prediction has always been calculated out.
3, because the domestic traffic is regularly: at work, Monday to 7 weeks this week also last week, about the same, so the prediction index or weight is the same.
4, holiday, take in the same period last year holiday weight data curve, combined with last week's data base, to predict the traffic flow of holiday tomorrow.
5, in short: the thought of algorithm, is calculated on a weekly forecast curve of the next week, with the amount of three weeks after the three days as a base to predict traffic flow of next week,
The holiday special treatment.

CodePudding user response:

Moderators, if the large data, the score so little, I also not to answer questions.
Without a sense of achievement,
There are more buckle door block chain block.