Finding Japanese flatmates with Ruby

  • Published on

  • View

  • Download

Embed Size (px)


<ul><li><p>Finding Japanese flatmates with Ruby</p><p>Andrew Grimm25 March 2015</p></li><li><p>About me Andrew Grimm Previously worked as a bioinformatician (analysing </p><p>biological data using IT)</p></li><li><p>Background</p><p> Learning Japanese since 2011 Want to find Japanese flatmates</p></li><li><p>Alternative approaches</p><p> Japanese-language classifieds (Nichigo press, JAMS TV) Just searching for Japanese in flatshare ads</p></li><li><p>Main approach</p><p> Download flatshare ads on Gumtree for Sydney area Check if any words match Japanese names</p></li><li><p>name.japanese?</p></li><li><p>Wikipedia</p><p> Tried using the names of articles in Japan-related categories</p><p> Results were too messy</p></li><li><p>ENAMDICT Dictionary of Japanese and non-Japanese names in </p><p>Japanese</p><p>Kanji</p><p>Hiragana</p><p>English translation</p><p>No kanji Katakana</p><p>English translation</p></li><li><p>ENAMDICT XML</p></li><li><p>Difficulties</p><p> No frequency data How often is a word used as a Japanese name? How often is a word used as a non-Japanese name? How often is it used as an ordinary word in English?</p></li><li><p>Example messy results</p><p> Large modern bathroom with japanese bathtub and shower in the apartment.</p><p> Indian, Bangladeshi, Italian people with Japanese-looking names</p><p> People looking for Japanese flatmates Machine is a Japanese name</p></li><li><p>Medline Database of scientific papers</p></li><li><p>Advantages of Medline</p><p> Should help with frequency data Can look at which names co-occur Look at author affiliation</p></li><li><p>Medline XML</p></li></ul>