Hi ,
I want to monitor traffic using this software. Basically i have data as shown below.
Let say we have 5 operator where we send traffic. I have generated a report based on date/hr,dayname,operator , # of message send in that hrs.
I have generated report this way because i have notice # of traffic send to operator depends on hrs and dayname. Ex Sat /Sun i have notice volume is low but on weekdays it high.Some particular hrs have 0 records.Some hrs have high traffic. Note this data are all dummy data.
Now what i want to do this suppose a job will be run at hh:05 hrs and it will check hh-1 hrs data mean last hrs data. If #of traffic send =0 then it will analysize data( all prev days/hrs may be last 50 days or more) and can send alert .
Currently we do some cron job.But logic is simple,hrly job run at hh:55 hrs. It check if hh hrs have=0 msg and prev hrs(hh-1) have >50 and prev day(current date-1) hh hrs have >50 records and 7 days early hh hrs also have >50 records then we send alert.
Some time alert doest fire if one of the condition fail. So operation try to adjust >50 to 25 and do some trail and error which i dont like.
That why i want use this software to predict traffic for better monitoring.
Which alogorithm will be best to use?
Thanks and regards
Sandip
date-hr | Dayname | operator | # of msg send |
2014-01-01 00 | Wednesday | Opr-1 | 24 |
2014-01-01 00 | Wednesday | Opr-2 | 5 |
2014-01-01 00 | Wednesday | Opr-3 | 78 |
2014-01-01 00 | Wednesday | Opr-4 | 0 |
2014-01-01 00 | Wednesday | Opr-5 | 45 |
2014-01-01 01 | Wednesday | Opr-1 | 89 |
2014-01-01 01 | Wednesday | Opr-2 | 23 |
2014-01-01 01 | Wednesday | Opr-3 | 43 |
2014-01-01 01 | Wednesday | Opr-4 | 3 |
2014-01-01 01 | Wednesday | Opr-5 | 4 |
contd… | |||
2014-01-01 23 | Thusrday | Opr-1 | 89 |
2014-01-01 23 | Thusrday | Opr-2 | 23 |
2014-01-01 23 | Thusrday | Opr-3 | 43 |
2014-01-01 23 | Thusrday | Opr-4 | 3 |
2014-01-01 23 | Thusrday | Opr-5 | 4 |
and so on upto 18 Feb current hrs |