noon99jaki/pra

Repository for the Path-Ranking Algorithm

pra's Issues

RelationLine.getDomains() no value

Do we need to add a file called relations to store query relations? Otherwise the Create Queries step only generates a createQueries.filter_errors file without content because RelationLine.getDomains() is empty.

The import edu.cmu.minorthird cannot be resolved

there is no package 'edu.cmu.minorthird', and lots of errors show up

Path filtering code in paper "lao2011kb"

Thanks first for opening source code for such a good work.

I want to use the path filtering strategies in paper Random walk inference and learning in a large scale knowledge base, which in section 2.3 Data-Driven Path Finding and 2.4 Low-Variance Sampling. Could you provide some instructions to find and use the code implementation of above methods? It's a little hard for me a new beginner to figure out the whole project structure :(

Thanks a lot again!

The output of prediction is empty

I set

task=predict

And run

java -cp pra-src-20140421.jar edu.cmu.pra.SmallJobs createQueries ../graphs/NELL165/ ./queriesR_train/ false false
java -cp pra-src-20140421.jar edu.cmu.pra.LearnerPRA

But the output in ./predictions are two empty files.
How should I do predictions?

---

I change to run

java -cp pra-src-20140421.jar edu.cmu.pra.SmallJobs createQueries ../graphs/NELL165/ ./queriesR_test/ false false
java -cp pra-src-20140421.jar edu.cmu.pra.LearnerPRA

It seems that I get the correct result. Thank you anyway!

The import edu.cmu.nlsa cannot be resolved

I just import this repository pra in Eclipse ,but it throw out this erro"The import edu.cmu.nlsa cannot be resolved",Can you help me to solve this question? where can i find the nlsa.jar?

where is the queriesR_test data?

where is the queriesR_test data?

thanks

http://www.cs.cmu.edu/~nlao/data/ can not be accessed.

Hi,
Now, http://www.cs.cmu.edu/~nlao/data/ seems inaccessible. So I don't know how to get the jar file and data files, can you tell me how to solve this problem?

Bias is ignored while loading models

During loading models, bias is ignored and set to zero due to the third parameter here is true. Is that normal?

pra/edu/cmu/lti/algorithm/optimization/AModel.java

Line 43 in cf02b24

for (VectorS vs: FFile.enuRows(model_file, "\t",true)){

Thanks!