data type in GeNIe

The front end.
Yang Yajie
Posts: 14
Joined: Thu Mar 19, 2020 11:49 am

data type in GeNIe

hello, I am using GeNIe to built the dynamic Bayesian belief network. I have a question about the value label in data columns.

such as I use 0,1,2,3,4,5,6,7,8,9 to represent age (0,10],(10,20], (20,30],...,(90,>90] in .csv file, then I input this csv file into GeNIe to make "learn structure" and "learn network'. I notice the dynamic bayesian network you provide on the website that we can use string to represent our value label. like True and False, Pittsburg and Saraha in the .csv file. if I use 1,2,3,.., 99 to represent the value label in the file, does this number has numeric meaning? like 2 is bigger than 1, 87 is bigger than 23? or does these number has the same equal meaning with "A,B,C,...,P'?

At first, I use the "A,B,C,...,S,M,L" as the value label, but when i check the correlation in GeNIe, these variables which represent by "A,B,C,..S,M,L" will not shown in the correlation table. So i change all this lable into ‘1,2,3,...11,12,13...’ instead of them in each varibale. and the correlation table will shown these variables. Because I want to check the correlations among these variables, So I am afraid I use the wrong value label in the CSV file and cause the wrong correlation. Could oyu please to help me with this problem?

The second question is I put discrete data and continuous data into one CSV file, when I make structure learning in GeNIe, it told me "Can't start learning due to problems with data. Selected algorithm supports continuous or discrete columns, but they can't be used together. ". Could you please tell me how to deal with this problem?

kind regards!
Yajie

shooltz[BayesFusion]
Posts: 1255
Joined: Mon Nov 26, 2007 5:51 pm

Re: data type in GeNIe

such as I use 0,1,2,3,4,5,6,7,8,9 to represent age (0,10],(10,20], (20,30],...,(90,>90] in .csv file, then I input this csv file into GeNIe to make "learn structure" and "learn network'. I notice the dynamic bayesian network you provide on the website that we can use string to represent our value label. like True and False, Pittsburg and Saraha in the .csv file. if I use 1,2,3,.., 99 to represent the value label in the file, does this number has numeric meaning? like 2 is bigger than 1, 87 is bigger than 23? or does these number has the same equal meaning with "A,B,C,...,P'?
If your data columns contains integers only, they will be sorted numerically (8 is less than 23, for example).
At first, I use the "A,B,C,...,S,M,L" as the value label, but when i check the correlation in GeNIe, these variables which represent by "A,B,C,..S,M,L" will not shown in the correlation table. So i change all this lable into ‘1,2,3,...11,12,13...’ instead of them in each varibale. and the correlation table will shown these variables. Because I want to check the correlations among these variables, So I am afraid I use the wrong value label in the CSV file and cause the wrong correlation

The second question is I put discrete data and continuous data into one CSV file, when I make structure learning in GeNIe, it told me "Can't start learning due to problems with data. Selected algorithm supports continuous or discrete columns, but they can't be used together. ". Could you please tell me how to deal with this problem
You should check either the discrete data, or continuous data before you proceed to learn the structure.

Yang Yajie
Posts: 14
Joined: Thu Mar 19, 2020 11:49 am

Re: data type in GeNIe

here is the data file. I use 1,2,3,..10 to represent different status. like gender 1=Male, 2=Female.
Thanks
Attachments
example.csv