# Headline aliases — maps a full name/phrase to a canonical short form so that # "Federal Reserve" and "the Fed" normalize to the same token and cluster together. # # Format: one entry per line under [aliases]: # "full phrase" = "canonical" # # Matching rules: # * Case-insensitive and after punctuation removal (so "people's" == "peoples"). # * Whole-word only — "united states" will not match inside "united statesman". # * Longest phrases are applied first — "president of the united states" wins # over "united states". # * The canonical form MUST NOT be a stopword (see resources/stopwords.txt), or # it will be stripped during normalization and the alias is lost. [aliases] # ---------------------------------------------------------------- US government "federal reserve" = "fed" "federal reserve system" = "fed" "white house" = "wh" "central intelligence agency" = "cia" "federal bureau of investigation" = "fbi" "national security agency" = "nsa" "department of justice" = "doj" "department of defense" = "dod" "department of homeland security" = "dhs" "environmental protection agency" = "epa" "securities and exchange commission" = "sec" "internal revenue service" = "irs" "social security administration" = "ssa" "centers for disease control" = "cdc" "centers for disease control and prevention" = "cdc" "food and drug administration" = "fda" "federal aviation administration" = "faa" "federal communications commission" = "fcc" "federal trade commission" = "ftc" "federal emergency management agency" = "fema" "national aeronautics and space administration" = "nasa" "national institutes of health" = "nih" "national transportation safety board" = "ntsb" "supreme court" = "scotus" "supreme court of the united states" = "scotus" "house of representatives" = "house" "president of the united states" = "potus" "vice president" = "vp" "prime minister" = "pm" "secretary of state" = "secstate" "attorney general" = "ag" # ---------------------------------------------------------- International bodies # "un" and "who" are stopwords, so these use concatenated forms to survive. "united nations" = "unitednations" "world health organization" = "worldhealthorg" "united nations security council" = "unsc" "international monetary fund" = "imf" "world trade organization" = "wto" "north atlantic treaty organization" = "nato" "international criminal court" = "icc" "international atomic energy agency" = "iaea" "organization of the petroleum exporting countries" = "opec" "european union" = "eu" "african union" = "au" "association of southeast asian nations" = "asean" # ----------------------------------------------------------- Countries & regions "united states" = "us" "united states of america" = "us" "united kingdom" = "uk" "great britain" = "uk" "united arab emirates" = "uae" "people's republic of china" = "china" "republic of korea" = "south korea" "democratic people's republic of korea" = "north korea" "saudi arabia" = "saudi" "middle east" = "mideast" "asia pacific" = "apac" "latin america" = "latam" # ----------------------------------------------------------- Politics & economy "republican party" = "gop" "democratic party" = "democrats" "gross domestic product" = "gdp" "gross national product" = "gnp" "consumer price index" = "cpi" "initial public offering" = "ipo" "unemployment rate" = "unemployment" "inflation rate" = "inflation" "national debt" = "debt" "budget deficit" = "deficit" # ---------------------------------------------------------- Technology & other "artificial intelligence" = "ai" "machine learning" = "ml" "electric vehicle" = "ev" "international space station" = "iss"