Gives different output when tested on other algorithm about spmf-py

[LoLei]

Interesting, thanks for reporting. spmf-py is largely untested for algorithms other than sequential pattern mining, such as association rule mining, although in theory the wrapper should work the same.

Could you please post the calls you make to both spmf-py and the jar itself, so I can reproduce it?

[Choneke]

I tried to run this on spmf-py

spmf = Spmf("TopKClassRules", input_filename="Input.txt",
output_filename="output.txt", arguments=[10, 0.05, 1, 101, 2])

[LoLei]

Could you also post how you call the jar itself?

[Choneke]

This is how jar is called in init.py file which is under spmf folder.

class Spmf:
    def __init__(self,
                 algorithm_name,
                 input_direct=None,
                 input_type="normal",
                 input_filename="",
                 output_filename="spmf-output.txt",
                 arguments=[],
                 spmf_bin_location_dir=".",
                 memory=0):
        self.executable_dir_ = spmf_bin_location_dir
        self.executable_ = "spmf.jar"

And below picture shows how the file is arranged, and I run example2.py to get my output

[LoLei]

Thanks, but I was thinking about this:

But when I run in jar file with the same input parameter the output which I expected is what I get.

Am I not understanding it correctly? That you also ran the spmf.jar file without spmf-py to get the correct output? I'm interested in that call.

Btw, one thing that seems suspicious is that in the documentation it says to run the jar like this: java -jar spmf.jar run TopKClassRules contextIGB.txt output.txt 7 80% 1,2, with 80% instead of 0.8, even though further below it says:

parameter minconf representing the minimum confidence that the association rules should have (a value in [0,1] representing a percentage).

Also I see in spmf-py, you're passing the "required items" parameter as multiple parameters. Try passing it differently. E.g. as one string, instead of separate parameters. I.e. like this: spmf = Spmf("TopKClassRules", input_filename="Input.txt", output_filename="output.txt", arguments=[10, 0.05, '1,101,2']).