I am having trouble when testing the Datasets with the corrupted one.

I have attached the file where you can view. As you can see at bottom there is a testing section where after I entered the corrupted value it should adjust it and give me back the nearest value. But here, it gives me higher value where it cant acceptable. guide me guys.

Answer 1

[Image Analyst]

It's giving you the maximum of the absolute deviation of the differences. This will always be a number >= 0, just like you told it. What do you think the value is higher THAN ? If you don't like the value you told it to give you, then what value would you prefer .

[Alagesan]

I want that corrupted value that I insert in testRecord column 5 to be like the original. e.g:

testRecord= [27,59,1725,35, 51 ];-------column 5= 51( this value I insert to test the data. If you looking the coding W represent a differential Hebbian rule]

after I run I want the column 5 to be exact like the original value or atleast nearest. So I can round up it.

original= [27 59 1725 35 55 ]------column 5 shows the original.

so after I run the testing part it must manage to adjust the corrupted value ]