I - Content ----------- This package contains files related to inference on the set of middles, as presented in the paper From C.H. Bryant and D. Fredouille "A Parser for the Efficient Induction of Biological Grammars". More precisely, this files concern Section 3 of this paper. This package contains: - A README file (this file) - example files for the inference process: middle.f middle.n - background knowledge files for the inference process middle_dftd.b middle_dftd_cost.b middle_yap.b middle_yap_cost.b - Prolog instruction files to start inference middle_dftd.pl middle_dftd_cost.pl middle_yap.pl middle_yap_cost.pl - csh scripts to start the inference processes middle_dftd.csh middle_dftd_cost.csh middle_yap.csh middle_yap_cost.csh I - Use of the package ---------------------- You need installed: - The Yap Prolog interpreter (http://www.ncc.up.pt/~vsc/Yap/yap.html) - The Aleph ILP system (http://web.comlab.ox.ac.uk/oucl/research/areas/machlearn/Aleph/aleph_toc.html) - The software presented in the paper "A Parser for the Efficient Induction of Biological Grammars". This software can be found at: http://www.comp.rgu.ac.uk/staff/chb/research/data_sets/ilp05/ Section 3 of the paper considers four variants on the inference process: - i-yap corresponding to files middle_yap.* - i-yap-cost corresponding to files middle_yap_cost.* - i-dftd corresponding to files middle_dftd.* - i-dftd-cost corresponding to files middle_dftd_cost.* To run one of these inference process: 1) Except for i-dftd, edit the corresponding .pl file and modify the first line to reflect the location of Aleph code on your machine. 2) For experiments using the dftd parser (middle_dftd and middle_dftd_cost), edit the corresponding .pl file and modify the lines downloading the DFTD parser to reflect the location of this parser on your machine: You have three lines to change: :- load_foreign_files(['../dftdp_in_Yap/c/grammar_parser'],[], download_cfgp_predicates). :- ['../dftdp_in_Yap/pl/grammar_parser_clause_to_grammar_rule.pl']. :- ['../dftdp_in_Yap/pl/grammar_parser_BK_to_C_structures.pl']. 3) Experiments of the paper considered different values for the "nodes" parameter of Aleph. To choose the value you want, edit the file .b corresponding to the algorithm you are interested in, and modify the set(nodes,value) line according to your wishes. 4) Run the script .csh corresponding to the inference process you are considering. Outputs: - a file xxx.rules, where xxx is either middle_yap, middle_yap_cost, middle_dftd or middle_dftd_cost depending on the inference process you consider. This file contains the set of inferred rules (which should be the same for all algorithms) - the screen where, after execution, the running time will be displayed. In case of errors during execution, they also should appear on the screen.