B3 コアタイム 第2回目 ( 2014年11月25日(火) )

Transcript

1. B3ίΞλΠϜ (2014೥11݄25೔(Ր)) ʰೖ໳ ࣗવݴޠॲཧʱ ɹɹ1ষ ݴޠॲཧͱPython ௕Ԭٕज़Պֶେֶ ࣗવݴޠॲཧݚڀࣨ B3 ழມ ܚथ
2. ςʔϚ
3. ςʔϚ ❖ PythonͷπʔϧͱςΫχοΫ ❖ NLTKͷ࢖༻ํ๏ ❖ ؆୯ͳ౷ܭॲཧ
4. Python
5. Python ❖ Pythonͷར఺ ❖ ߏจ΍ҙຯͷཧղ͕ฏқ ❖ จࣈྻΛऔΓѻ͏ؔ਺͕๛෋ ❖ Pythonͷಛ௃ ❖ ม਺ͷܕͷએݴ͕ෆཁ ❖ ߏจʹΠϯσϯτΛ࢖༻͢Δ https://www.python.org/community/logos/
6. NLTK
7. NLTK ❖ ඼ࢺͷλά෇͚, ߏจղੳ, ςΩε τͷ෼ྨͱ͍ͬͨॲཧΛߦ͏Πϯ λʔϑΣΠε ❖ ෳࡶͳ໰୊Λղͨ͘Ίʹ૊Έ߹Θ ͤΔ͜ͱ͕Ͱ͖Δ࣮૷ Natural Language Toolkit
8. ࢖༻͢Δલʹ ❖ NLTKͷΠϯϙʔτ ❖ >>> import nltk ❖ ςΩετͷΠϯϙʔτ ❖ >>> from nltk.book import *
9. NLTKͷػೳ ❖ ςΩετͷݕࡧ ❖ >>> text.concordance(“hoge1”) ❖ >>> text.concordance([“hoge1”, “hoge2”]) ❖ ୯ޠΛ਺͑Δ ❖ >>> text.count(“hoge”)
10. NLTKͷػೳ ❖ ޠኮΛ਺͑Δ ❖ >>> len(text) ❖ ޠኮͷऔಘ ❖ >>> set(text) ❖ ޠኮͷιʔτ ❖ >>> sorted(set(text))
11. NLTKͷػೳ ❖ ؔ਺ͷఆٛ ❖ ྫ) ςΩετͷޠኮͷ๛෋͞ ❖ >>> def function(text) : ❖ . . . return len(text) / len(set(text))
12. Python ͖ͭͮ
13. Pythonͷػೳ ❖ Ϧετ ❖ ςΩετͳͲΛऩೲ͢Δ഑ྻͷΑ͏ͳ΋ͷ. ❖ >>> sent[“hoge1”, “hoge2”, “hoge3”, “hoge4”]
14. Ϧετ ❖ Ϧετͷ࿈݁ ❖ >>> sent1[“hoge1”, “hoge2”] + sent2[“hoge3”, “hoge4”] ❖ Ϧετͷ௥Ճ ❖ >>> sent.append(“hoge”) ❖ εϥΠγϯά ❖ >>> sent[5 : 8]
15. Pythonͷػೳ ❖ จࣈྻ ❖ Ճࢉ, ৐ࢉͳͲ΋Մೳ. ࿈݁, ෼ׂͷػೳΛ࣋ͭ. ❖ name = ‘Hoge’
16. จࣈྻ ❖ จࣈྻͷ࿈݁ ❖ >>> ‘ ‘ .join([‘Hoge1’, ‘Hoge2’]) ❖ จࣈྻͷ෼ׂ ❖ >>> ‘Hoge1 Hoge2’.split()
17. Pythonͷػೳ ❖ ৚݅ࣜ ❖ ༻๏͸Cݴޠͱ΄΅ಉ༷. ൣғΛΠϯσϯτͰ൑அ. ❖ >>> if len(word) < 5 ❖ . . . print ‘word length is less then 5’
18. ৚݅ࣜ ❖ ୯ޠൺֱʹ࢖͑Δԋࢉࢠ ❖ s.startswitch(t) s ͕ t Ͱ࢝·Δ͔Ͳ͏͔ ❖ s.endswitch(t) s ͕ t ͰऴΘΔ͔Ͳ͏͔ ❖ t in sɹɹɹɹɹ s ͷதʹ t ؚ͕·ΕΔ͔Ͳ͏͔ ❖ s.islower() s ͷதʹେจࣈؚ͕·Ε͍ͯͳ͍ ❖ s.isupper() s ͷதʹࢠจࣈؚ͕·Ε͍ͯͳ͍ ❖ s.isalpha() s ͷதͷจࣈ͕શͯΞϧϑΝϕοτ͔Ͳ͏͔ ❖ s.isalnum() s ͷதͷจࣈ͕શͯΞϧϑΝϕοτ΋͘͠͸਺ࣈ͔Ͳ͏͔
19. ৚݅ࣜ ❖ ୯ޠൺֱʹ࢖͑Δԋࢉࢠ ͖ͭͮ ❖ s.isdigit() s ͷதͷจࣈ͕શͯ਺ࣈ͔Ͳ͏͔ ❖ s.istitle() s ͕λΠτϧέʔε͔Ͳ͏͔
20. Pythonͷػೳ ❖ ৚݅෼ذ ❖ ༻๏͸Cݴޠͱ΄΅ಉ༷. ಉ͘͡ൣғΛΠϯσϯτͰ൑அ. ❖ >>> for i in range(10) ❖ . . . print i
21. ؆୯ͳ౷ܭॲཧ
22. ؆୯ͳ౷ܭॲཧ ❖ සग़͢Δ୯ޠΛऔΓग़͢ ❖ >>> fdist1 = FreqDist(text) ❖ ςΩετதͷҟͳΓޠΛऔΓग़͢ ❖ >>> vocaburaly1 = fdist1.keys()
23. ؆୯ͳ౷ܭॲཧ ❖ ςΩετͷ಺༰Λද͢୯ޠΛऔΓग़͍ͨ͠. ❖ → औΓग़ͨ͠୯ޠ͔ΒςΩετͷ಺༰͕ཧղͰ͖Δʂ
24. ؆୯ͳ౷ܭॲཧ ❖ ‘,’ ΍ ‘the’, ‘.’, ‘and’ͳͲจࣈ਺͕୹͍୯ޠ͕සग़͢ΔͳΒ͹, 1ճ͠ ͔ొ৔͠ͳ͍୯ޠΛ୳ͤ͹Α͍ͷͰ͸ʁ ❖ 15จࣈҎ্ͷ୯ޠΛऔΓग़͢ ❖ >>> V = set(text1) ❖ . . . long_words = [w for w in V if len(w) > 15]
25. ؆୯ͳ౷ܭॲཧ ❖ 1ճ͔͠ొ৔͠ͳ͍୯ޠ͸,จ຺͔Βཧղ͠ͳ͚Ε͹ͳΒͳ͍୯ޠ͕ ଟ͘, ςΩετͷ಺༰͸ཧղͰ͖ͳ͍. ❖ 7จࣈҎ্, 7ճҎ্ొ৔ͨ͠୯ޠΛऔΓग़͢ ❖ >>> fdist5 = FreqDist(text5) ❖ >>> sorted([w for w in set(text5) if len > 7 and fdist5[w] > 7])