Predicting delays in a multi node supply chain of a courier company?


Hello Everyone…

I am currently working on a project which requires me to predict if there might be an delay in delivery of the consignments. Any particular consignment would have a lane decided. per say if an consignment is supposed to go from Delhi to Chennai. It might pass thought various hubs.
sender -> Pickuphub(at delhi) -> mother hub (at Bhopal) -> delivery hub (chennai) -> receiver.

Lets say there are about 25 pickup hubs, 16 mother hubs and 30 delivery hubs. So the total number of combinations would be 25 X 16 X 30. any time further than the promised time is considered as a delay.

So can i use logistic regression for finding true/false for the predicting the delay. I m using xlminer on excel to test the data.


You could also use linear regression on the time taken. Then, predict delay as true if the time taken is greater than the expected time. This approach could work better in my opinion.


All the predictive factors are categorical in nature and there are no numerical values. Can i use logistic regression and SVM on them>??