Test Lead-End-Unit in Exclusive Filter

I. Introduction

This section describes the testing processes and results of the exclusive filters of nonLead and nonEnd units. The Lexicon multiwords are used to tested. Ideally, the results should not filter out any multiwords (or min.) in Lexicon.

TBD:

• This section of Java codes are moved to TBD because some files are under developement. We will need to come back to complete this sectiondue to the time constraint.

II. Processes

• directory: ${MULTIWORDS_DIR}/bin
• program: 3.NonLeadEndTerm
• Run program: shell> ./3.NonLeadEndTerm ${YEAR}
• Processes:

  Step	Description	IO	Notes - Examples
  15	Get ruleType on multiwords from Lexicon GetLetRtForTermsInLexicon.java Assign invalid Lead-End-Unit ruleTypes on Lexicon multiwords: RT_INV_LEAD_TERM RT_INV_END_TERM RT_INV_END_ABB RT_INV_LEAD_END_TERM	Inputs: ./outData/3.InvalidLeadEndTerm/lexMultiwords.data Outputs: ./outData/3.InvalidLeadEndTerm/lexMultiwords.data.ruleType ./outData/3.InvalidLeadEndTerm/lexMultiwords.data.ruleType.ilet (10)	1 min. Only 10 exceptions, all of them are RT_INV_END_ABB => Algorithm of endWord with abbreviation pattern can be improved
  16	Analyze ruleType on multiwords from Lexicon AnalyzeLetRtForTermsInLexicon.java Analyze results from above step (10) Get the precision of exclusive fitler on Lexicon	Inputs: ./3.InvalidLeadEndTerm/lexMultiwords.data.ruleType ./3.InvalidLeadEndTerm/lexMultiwords.data.ruleType.exceptions outputs: ./3.InvalidLeadEndTerm/lexMultiwords.data.ruleType.rpt	5 sec. precision: 99.9981% 1 invalid ruleType: RT_INV_END_ABB
  17	Get multiwords in Lexicon by lead/end units GetLexiconMultiwordsByLeadEndTerm.java Find all multiwords in Lexicon by specifying lead/end word	Inputs: ./outData/3.InvalidLeadEndTerm/lexMultiwords.data.ruleType Outputs: ./outData/3.InvalidLeadEndTerm/LexiconMw/lexMultiwords.data.ruleType.${LEAD_END_WORD}	5 sec. Used for case study