Shortest Routes

_maden_

Hello! I want to schedule and find the shortest routes of ferry ways. For that, I made an Excel file, in which there are a route beginning column, so called 'from' , a route end column 'to' and time columns. For example I want to go from 'A' to 'C'. To be able to do this, I have to go to 'A' from 'B' and then from 'B' to 'C'. So, normally I think that I have to write a code like "if from==to" but I have to do this with rapidminer. I also have to find the proper time for the routes from the Excel file. So, how can I find these relationships of 'route beginning' and 'route end' by taking time factor also in consideration?
I would appreciate any help!

lionelderkrikor

Hi @_maden_,

Your use case seems an interesting problem but first I have to quote Seneque
(a Stoic school philosopher, playwright, and first-century Roman statesman)., a quote particularly suited to your case : 

"Long is the road by precept, short and easy by example."


So could you give an example of what you have in entry and an example of what you want to obtain ?

Regards,

Lionel

_maden_

Hi @lionelderkrikor ,

Thank you for your comment. 

There is, for example, an 'A' station and from this station I want to go to 'D' station. There are ways like 'A'>'D' or 'A'>'B'>'D' or 'A'>'C'>'D' etc. So I try to find out what the shortest way is, by taking time factor also in consideration. A is let's say an element of first column an D is an element of second. I want to find out the other relationships to this two rows, so that I can see the routing options and times. 

Telcontar120

This is going to be a problem that is tricky for RapidMiner to solve, at least not without a lot of data restructuring.  ML algorithms are designed to take each row (example in RapidMiner) as a unique case and then take a specific value within that row as the target (label in RapidMiner) to predict.
It sounds like you have some raw data structure which requires you to interpret which of many multiple paths are feasible from one destination to another.  But what are the rules that enable you to determine that?  How could they be programmed or represented for a non-human?
You could try restructuring your data to show the connections between every pairwise combination of destinations, followed by every 3-set combination, ,etc., with their associated times.  But to do that is to solve the problem you are asking about!  So I am not sure how RapidMiner could best be used to accomplish this task.