I have a dataset, which contains measurement values of a mechanical process. The data is being recorded in 1 sec time interval. The data set contains, recorded measurement value of the process, Lower bound value and upper bound value and time stamp for three days.
The idea is to look at the pattern of the data and then identify which point could be reason in future that the product will fail.
To prove that statistically, I plotted the box plot and identified the outlier values. These outliers could be the reason that the products are failing in future.
Also, I need to mention the best operating region for that process.
I want to model, in such a way that when there are new measurement values in future,I should be able to recognize the pattern and find the point which could be the reason for fault.
Could any one help me , how I could proceed with this problem ? I am new to statistics and would be helpful, if you could provide me some idea on how to deal with it.