Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
B3_Seminar_04
Search
kakubari
February 16, 2017
Technology
0
71
B3_Seminar_04
長岡技術科学大学 自然言語処理研究室
角張竜晴
kakubari
February 16, 2017
Tweet
Share
More Decks by kakubari
See All by kakubari
動詞クエリの語間の関係性に基づくクエリマイニング
kakubari
0
110
Neural Modeling of Multi-Predicate Interactions for Japanese Predicate Argument Structure Analysis
kakubari
1
160
Leveraging Crowdsourcing for Paraphrase Recognition
kakubari
0
83
Automatically Acquired Lexical Knowledge Improves Japanese Joint Morphological and Dependency Analysis
kakubari
0
100
Labeling the Semantic Roles of Commas
kakubari
0
79
Integrating Case Frame into Japanese to Chinese Hierarchical Phrase-based Translation Model
kakubari
0
120
Improving Chinese Semantic Role Labelingusing High-quality Surface and Deep Case Frames
kakubari
0
90
Exploring Verb Frames for Sentence Simplification in Hindi
kakubari
0
130
述語項構造と照応関係のアノテーション
kakubari
0
230
Other Decks in Technology
See All in Technology
AI時代に必要なデータプラットフォームの要件とは by @Kazaneya_PR / 20251107
kazaneya
PRO
4
680
Boxを“使われる場”にする統制と自動化の仕組み
demaecan
0
200
SOTA競争から人間を超える画像認識へ
shinya7y
0
690
MCP サーバーの基礎から実践レベルの知識まで
azukiazusa1
18
8.2k
InsightX 会社説明資料/ Company deck
insightx
0
200
Amazon Athena で JSON・Parquet・Iceberg のデータを検索し、性能を比較してみた
shigeruoda
1
300
現場の壁を乗り越えて、 「計装注入」が拓く オブザーバビリティ / Beyond the Field Barriers: Instrumentation Injection and the Future of Observability
aoto
PRO
1
940
ソフトウェア品質を支える テストとレビュー再考 / 吉澤 智美さん
findy_eventslides
0
110
Data Engineering Guide 2025 #data_summit_findy by @Kazaneya_PR / 20251106
kazaneya
PRO
8
1.3k
技術の総合格闘技!?AIインフラの現在と未来。
ebiken
PRO
0
130
OPENLOGI Company Profile for engineer
hr01
1
46k
次世代のメールプロトコルの斜め読み
hirachan
3
380
Featured
See All Featured
No one is an island. Learnings from fostering a developers community.
thoeni
21
3.5k
Why You Should Never Use an ORM
jnunemaker
PRO
60
9.6k
It's Worth the Effort
3n
187
28k
Practical Tips for Bootstrapping Information Extraction Pipelines
honnibal
PRO
23
1.5k
Templates, Plugins, & Blocks: Oh My! Creating the theme that thinks of everything
marktimemedia
31
2.6k
For a Future-Friendly Web
brad_frost
180
10k
[RailsConf 2023] Rails as a piece of cake
palkan
57
6k
The Cult of Friendly URLs
andyhume
79
6.7k
Building an army of robots
kneath
306
46k
Principles of Awesome APIs and How to Build Them.
keavy
127
17k
Producing Creativity
orderedlist
PRO
348
40k
Building Better People: How to give real-time feedback that sticks.
wjessup
370
20k
Transcript
Ԭٕज़Պֶେֶ ిؾిࢠใֶ՝ఔ ֶ෦ɹ֯ுཽ ࣗવݴޠݚڀࣨ ɹ#̏θϛ ʙୈճʙ ϏοΫσʔλղੳೖᶄ
目次 ˔౷ܭͷجૅ ˔֬ີؔɾྦྷੵؔ
統計の基礎 ˔ఆৗͱ ɹ࣌ܥྻղੳͰඞཁͱͳͬͯ͘Δ֓೦ ˔࣌ܥྻͱ ɹ࣌ؒͷྲྀΕͱڞʹ؍ଌྔͷมԽ͕ه͞Εͨσʔλ ɹྫ͑ɾɾɾ ɹɾҝସגͷՁ֨ ɹɾಉ͡ॴͷؾԹؾѹ
統計の基礎 ͋Δ࣌ࠁ̓ʹ؍ଌ͞ΕͨΛ͇ ̓ Ͱද͢ɻ ࣌ܥྻ͕ఆৗͰ͋ΔͨΊͷ݅ɾɾɾ ɾฏۉ͕࣌ؒʹΑΒͣҰఆɹ ɾࢄ͕࣌ؒʹΑΒͣҰఆ ɾࣗݾڞࢄ͕࣌ؒࠩͷؔ ͜͜ͰɺЖ
Мఆɺ̺࣌ؒࠩΛද͢ɻ E[x(t)] = µ E[x(t)− µ]2 = σ 2 E[(x(t)−µ)(x(t − k)−µ)]= C(k)
統計の基礎 ˔౷ܭղੳΛߦ͏্Ͱఆৗੑɺඇৗʹॏཁ ղੳΛ͢Δσʔλͷൣғ͕มΘͬͯಉ͡౷ܭ݁Ռ ͕ಘΒΕΔΛ͍ࣔࠦͯ͠Δɻ ͭ·Γɺ౷ܭ݁ՌʹൣғબʹΑΔۮવੑ͕བྷΉ͜ͱ Λഉআͯ͘͠ΕΔɻ
統計の基礎 ˔౷ܭղੳͱ ɹશମ͔Βൈ͖ग़ͨ͠Ұ෦ΛݟͯɺશମΛΔ ྫ͑ɾɾɾʮຊதͷখֶੜͷମॏΛௐࠪ͢Δʯ ɹௐࠪରɿຊશࠃͷখֶੜશһ ௐࠪͷରͱͳΔूஂΛूஂͱ͍͏ɻ ཧͱͯ͠ɺूஂΛͯ͢ௐࠪ͢ΕΑ͍ɻ
શௐࠪ
統計の基礎 ͕ͩɺूஂ͕େ͖͘ɺௐ͕ࠪࠔͰ͋Δɻ Ὃ ूஂ͔Β̽ݸΛൈ͖ग़ͯ͠؍ଌ͠ɺ ͔ͦ͜ΒશମͷಛΛਪఆ͢Δɻ
؍ଌͷूஂඪຊͱݺͿɻ ಛʹɺཁૉ͕̽ͷ߹େ͖̽͞ͷඪຊͱݺͿɻ
統計の基礎 ʙ౷ܭղੳΛߦ͏্Ͱॏཁͳ๏ଇʙ ˔େͷ๏ଇ ʮ͋Δूஂ͔Βແ࡞ҝநग़͞ΕͨඪຊฏۉඪຊͷαΠζΛ େ͖͘͢Δͱਅͷฏۉʢूஂͷฏۉʣʹۙͮ͘ʯ ˔த৺ۃݶఆཧ ʮฏۉЖɺࢄМΛ࣋ͭҙͷʹै͏ूஂ͔Βɺ େ͖̽͞ͷඪຊΛநग़ͨ࣌͠ɺඪຊฏۉ̚<͇>ͷɺ͕̽े େ͖͚ΕฏۉЖɺࢄМ̽ͷਖ਼نʹۙͮ͘ʯ
˔֬ີؔ ɹ֬มʢཧྔʣ̭͕ඍখͳ۠ؒ ʹͦͷΛͱΔ֬ʢ֬ີʣΛ༩͑Δؔ ɹ֬ม̭͕ɹɹɹɹɹͱͳΔ֬Λ ͱ͢Δͱɺ ֬ਖ਼Ͱ͋Γɺͦͷ͕̍Ͱ͋Δ͜ͱ͔Βɺ f
(x) x < X < x +δx P(a < X < b) = f (x)dx a b ∫ a < X < b P(a < X < b) f (x)dx −∞ +∞ ∫ =1 f (x) ≥ 0
確率密度関数・累積分布関数 ˔ώετάϥϜ ɹ۠ؒͷදΛԣ࣠ʹɺͦͷ۠ؒͷΛॎ࣠ʹͱͬͯࢹ֮Խ ͨ͠ͷ ֬มͷಛ͕Θ͔Δ ɹɾͲͷ͘Β͍͕ΓΛ͔࣋ͭ ɹɾҰ൪ଟ͍Կ͔
ͳͲʜ ώετάϥϜͷ֓ܗΛ͑ΔͨΊʹن֨ԽΛߦ͏ɻ ɹɾΛσʔλͰׂΓɺ֤۠ؒͰͷ֬Λܭࢉ͢Δ ɹɾ֤۠ؒͷ֬Λ۠ؒͷ෯ͰׂΓɺ֬ີΛܭࢉ͢Δ ɹɾԣ࣠ʹ֤۠ؒͷදɺॎ࣠ʹ֬ີΛϓϩοτ͢Δ
確率密度関数・累積分布関数 ˔ྦྷੵؔ ɹ֬มͷ͕͇ΑΓେ͖͘ͳΔ֬Λ༩͑Δؔ ֬ີؔΛ༻͍ͯɺ ͱఆٛ͞ΕΔؔ'Λ֬ม̭ͷྦྷੵؔͱ͍ ͏ɻ F(x)
= P(X > x) = f ( ! x )d ! x x ∞ ∫
確率密度関数・累積分布関数 ˔ར ۠ؒΛ۠Δඞཁ͕ͳ͍ͨΊɺσʔλ͕ൺֱతগͳ͘ ͯ͋Δఔ͖Ε͍ʹඳ͚Δɻ σʔλͱྦྷੵؔ̍ର̍ʹରԠ͢Δɻ ˔άϥϑͷॻ͖ํ ɹ֬ີؔʢੵʣΛܦ༝͢Δํ๏ɺݫີͳ݁Ռ
Λಘ͍ͨ߹ʹ΄ͱΜͲΘΕͳ͍ɻ࣮ࡍɺσʔλͷ ιʔτͰٻΊΔɻ ᾇ̣ݸͷσʔλΛେ͖͍ॱʹฒΔ ᾈঢॱʹσʔλʹରͯ͠ɺ͔̍Β/·ͰॱҐ3Λ͚ͭΔ ᾉσʔλͷΛԣ࣠ʹɺ3/Λॎ࣠ʹϓϩοτ͢Δ
参考文献 ˔ߴ҆ඒࠤࢠฤஶɺాଜޫଠɾࡾӜߤஶɺ ɹʮֶੜɾٕज़ऀͷͨΊͷϏοΫσʔλղੳೖʯ ʢୈ̍ষʙୈ̏ষʣɺ ɹגࣜձࣾຊධࣾɺ݄