Hi all,
we are having a data for 3 years in a column. we want to know the next dates in the same column. can anyone please suggest me how to do this? Hereby is the sample data.
SUBMIT_DATE | LAST_RESOLVED_DATE | MTTR_MINS | DESCRIPTION |
11/5/2016 15:52 | 11/9/2016 6:44 | 5220 | System: CPU Utilization is 99.03% busy - User: 60.14% System |
5/26/2017 16:16 | 5/31/2017 18:54 | 7380 | System: CPU Utilization is 99.59% busy - User: 99.24% System |
9/27/2016 21:08 | 10/2/2016 3:39 | 6180 | System: CPU Utilization is 99.97% busy - User: 3.10% System |
9/29/2016 18:24 | 10/2/2016 14:46 | 4080 | System: CPU Utilization is 99.80% busy - User: 98.92% System |
9/22/2016 19:31 | 9/25/2016 15:29 | 4080 | System: CPU Utilization is 99.90% busy - User: 98.89% System |
You are looking for a Predictive Model.
To predict the next date, you need to define the predictor variables, which columns should be used to predict the date column?
Do you have any experience working with predictive modeling?
5 points are not enough for any forecasting model. If you had several thousand you could try to use a time series forecast like this.
1) Use a Generate Rows tool to generate all valid dates for the last 5 years .
2) Join this data to your input and create a new field called event that is 1 if the event occurred on that date and 0 if it didn't.
3) Pass this to an ets tool with event as the dependent variable and look at the results.
Even once you do this, there's no guarantee of success. This is due to the domain of the problem. High server utilization is very rarely based on just time. It's generally caused by other factors, number of processes running, peak process utilization, number of concurrent users, etc. These are factors that are not captured in your input data set, so even if time series analysis can find a pattern, you'll have no understanding of the cause of the problem
Dan