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
ANNとナイーブベイズを使った雑な野球選手の成績予測 / Baseball player performance prediction with Python
Search
Shinichi Nakagawa
July 22, 2020
Research
0
2.8k
ANNとナイーブベイズを使った雑な野球選手の成績予測 / Baseball player performance prediction with Python
PyCon JP 2020で話す予定の話のダイジェストです.
kawasaki.rb #86 での練習試合.
#Python #DataScience #MLB #Baseball
Shinichi Nakagawa
July 22, 2020
Tweet
Share
More Decks by Shinichi Nakagawa
See All by Shinichi Nakagawa
Terraform, GitHub Actions, Cloud Buildでデータ基盤をProvisioningする / Data Platform provisioning for Google Cloud and Terraform
shinyorke
2
2.1k
Cloud RunとCloud PubSubでサーバレスなデータ基盤2024 with Terraform / Cloud Run and PubSub with Terraform
shinyorke
8
2.1k
自らを強いエンジニアにするための3つの習慣 / I need to be myself, I can't be no one else
shinyorke
76
53k
阪神タイガース優勝のひみつ - Pythonでシュッと調べた件 / SABRmetrics for Python
shinyorke
1
980
Pythonとクラウドと野球の推し活. / Baseball Data Platform for Python and Google Cloud
shinyorke
2
2.3k
月額コーヒー3.34杯分のコストでオオタニサンの活躍を見守るデータ基盤のはなし / Pyhack Con
shinyorke
2
390
俺のDXを実現するためのサーバレスなデータ基盤開発と運用 / Serverless Data Platform and Baseball
shinyorke
5
11k
機械学習エンジニアが目指すキャリアパスとその実話 / My Journey to Become a ML Engineer
shinyorke
6
14k
一人でも小さく始められるGoogle Cloudで実現するほぼサーバレスなデータ基盤 / Serverless Dataplatform for Google Cloud
shinyorke
0
440
Other Decks in Research
See All in Research
Evolutionary Optimization ofModel Merging Recipes (2024/04/17, NLPコロキウム)
iwiwi
9
3.2k
論文紹介 DISN: Deep Implicit Surface Network for High quality Single-view 3D Reconstruction / DISN: Deep Implicit Surface Network for High quality Single-view 3D Reconstruction
nttcom
0
120
デフスポーツにおける支援技術 〜競技特性・ルールと技術との関係〜
slab
0
220
How to Perform Manual Classification for Deep Learning Using CloudCompare
kentaitakura
0
670
F0に基づいて伸縮された画像文字からの音声合成 [ASJ2024春]
nehi0615
0
120
Alternative Photographic Processes Reimagined: The Role of Digital Technology in Revitalizing Classic Printing Techniques【SIGGRAPH Asia 2023】
toremolo72
0
440
LLMマルチエージェントを俯瞰する
masatoto
26
16k
Ground Metric Learning with applications in genomics
gpeyre
0
370
SANER 2019 Most Influential Paper Talk
tsantalis
0
120
[2023 CCSE] ZOZOTOWN検索における 研究開発の取り組みについて
tomoyayama
0
130
媒介分析と疫学
kingqwert
0
100
DeepCrysTet: A Deep Learning Approach Using Tetrahedral Mesh for Predicting Properties of Crystalline Materials
tsurubee
0
380
Featured
See All Featured
WebSockets: Embracing the real-time Web
robhawkes
59
7k
Fashionably flexible responsive web design (full day workshop)
malarkey
398
65k
Building Applications with DynamoDB
mza
88
5.6k
Designing on Purpose - Digital PM Summit 2013
jponch
111
6.5k
Build The Right Thing And Hit Your Dates
maggiecrowley
25
2k
Designing for humans not robots
tammielis
248
25k
Rebuilding a faster, lazier Slack
samanthasiow
74
8.2k
Large-scale JavaScript Application Architecture
addyosmani
504
110k
Writing Fast Ruby
sferik
622
60k
Building a Scalable Design System with Sketch
lauravandoore
457
32k
Bash Introduction
62gerente
605
210k
VelocityConf: Rendering Performance Case Studies
addyosmani
321
23k
Transcript
ٿબखͷ༧ଌϞσϧΛ ͍͍ײ͡ʹ࡞ͬͯΈͨVer 1.0 Shinichi Nakagawa (@shinyorke) kawasaki.rb #86 7पͪΐͬͱLTେձ
Who am I ? • Shinichi Nakagawa(@shinyorke) • JX௨৴ࣾγχΞɾΤϯδχΞ •
࠷ۙͣͬͱσʔλج൫ɾσʔλੳ͍ͯ͠ΔϚϯ • ຊདྷٿσʔλαΠΤϯεʹڧ͍ਓ • ٕज़ސ͡Ί·ͨ͠
kwsk.pyҊ݅Ͱ͢ :bow: PyCon JPʹ2ͿΓ6ճͷ⽁Λ͢Δ͜ͱʹͳΓ·ͯ͠. ٱ͠ͿΓʹนଧͪʹͬͯ·͍Γ·͓͖ͨ͠߹͍͍ͩ͘͞⽁
ʲਤʳࠓճ͖ͬͯͨ͜ͱ ຊ֨తͳ։ൃ4݄͔Β, ࠷ޙͷλεΫ͕௨ͬͨͷ͕͍ͭ࠷ۙ اըɾߏؚΊΔͱ࣮͍ۙϓϩδΣΫτͩͬͨΓ
None
σʔληοτ࡞ɾಛྔநग़ • ϝδϟʔϦʔάͷσʔλʮSean Lahmanʯʮretrosheetʯ ͜ΕΒΛͯ͢BigQueryʹimport • CSV͔Βςʔϒϧ࡞ • ػցֶशλεΫʹඞཁͳಛྔΛ۪ʹࢉग़
ػցֶशλεΫͦͷᶃ ʮࣅ͍ͯΔબखΫϥελΛ࡞Δʯ
कඋҐஔɾͷงғؾͰΫϥελϦϯά • ࡶʹݴ͏ͱ, ʮ˓˓ͬΆ͍બखϥϯΩϯάʯΛ࡞Δ • ྫ͑ࡔຊ༐ਓʢڊਓʣͬΆ͍બखʁͱݴΘΕͨΒ, ʮकඋҐஔ͕γϣʔτʯʮৗʹ3ׂ20ຊྥଧଧͭʯ ͱ͔ͦΜͳײ͡. γϣʔτͰ͋Δ͜ͱϚετ, ͋ͱଧܸ࣍ୈ.
• ଧܸ͓ΑͼҰ෦ͷकඋࢦඪΛͬͯϢʔΫϦουڑΛ ٻΊͯ૯ΓͰ֤બखͷʮͦΕͬΆ͍ϥϯΩϯάʯΛ࡞Εͦ͏.
ۙࣅ࠷ۙ୳ࡧʢANNʣͰͬͯΈͨ • kNNͱ͔k-meansͱ͔Γํ৭ʑ͚͋ͬͨͲANNͰͬͨ݁Ռ ͕͍͖ͳΓ͍͍ײͩͬͨ͡ͷͰ͜Εʹͨ͠. • ANNͷλεΫAnnoy͍ͬͯ͏ϥΠϒϥϦͰര։ൃ. • ϝδϟʔϦʔΨʔ19,000ਓͷσʔλͰͬͨΒ͍͍ײ͡ʹ.
ίʔυʢҰ෦ൈਮʣ˞ಛྔൿີ ֶश͔ΒϞσϧอଘͨͬͨ͜Ε͚ͩ. σʔλେ͖͘ͳ͍ͷͰඵͰऴΘΓ·ͨ͠.
ϚοτɾνϟοϓϚϯʢMLBएखࡾྥखʣʹ͍ۙબख ٬؍తͳσʔλ͔Β, ϑΝϯͱͯͬͯ͠Δͱͯ͠. ͍ۙબख͕ͪΌΜͱू·Γ·ͨ͠, શһࡾྥखͰଧܸͰ݁Ռग़ͤΔϚϯͳͷͰจ۟ͳ͠ʂ
ࣅ͍ͯΔબखूΊʹޭ ʢଞͷϙδγϣϯ͍͍ײͩͬͨ͡ʣ ޙ͔ͬ͜ΒߋʹΧςΰϦʔྨͯ͠ ʮະདྷͷΛ࡞ΓࠐΉʯ ࣄ͕Ͱ͖ͨΒʂ
ػցֶशλεΫͦͷᶄ ʮಉ͡ΧςΰϦͷબखΛݟ͚ͭΔʯ
φΠʔϒϕΠζʹΑΔΧςΰϦʔ͚ • ࣗવݴޠॲཧͷྨλεΫΈ͍ͨͳղ͖ํͰͬͯΈͨ. • ީิʮφΠʔϒϕΠζʯʮϥϯμϜϑΥϨετʯ͋ͨΓ. ࠓճφΠʔϒϕΠζͰͬͨ. • ٿʹ͓͚Δ౷߹తͳೳྗࢦඪʮOPSʯΛ͝ͱͷΧςΰϦʔʹ͚, ͍͔ͭ͘ͷଧܸࢦඪΛϕΫτϧʹ࣮ͯ͠ࢪ. •
࣮ී௨ʹscikit-learnͱPandasͰΓ·ͨ͠.
ͬͨ͜ͱʢཁʣɹ˞ࡶʹॻ͍ͯ·͢ • ֶशσʔλ • ༧ଌ͍ͨ͠બखʹࣅͨબख50ਓͷΛϐοΫΞοϓ • ಛྔൿີͰ͕͢…ී௨ͷଧܸʹӅ͠ຯগʑ • ༧ଌσʔλ •
༧ଌ͍ͨ͠બखͷಛྔ • ݁Ռͷϥϕϧσʔλ • OPSΛ5ஈ֊ͷΧςΰϦʹͨ͠ͷ(1ʙ5) • ্هͰࢦఆͨ͠ΧςΰϦʹଐ͢Δબखͷྸผฏۉ͔ΒͦΕͬΆ͍Λग़͢
༧ଌͱҰॹʹݟͯΈ·͠ΐ͏͔.
ϚοτɾνϟοϓϚϯʢݱ࣮ͷʣ 24ʙ26ࡀʢڈ·Ͱʣͷ. ༧ଌ͍ͨ͠ͷ27ʙ29ࡀͷ.
ϚοτɾνϟοϓϚϯʢ༧ଌ͖ʣ 27ࡀҎ߱ͷΛ༧ଌͨ݁͠ՌΛؚΊͨάϥϑ.
None
ग़͖ͯͨ݁ՌΛ͡Δͱ… • ൺֱత, ݱ࣮ʹଈͯ͠ΔͬΆ͍݁ՌʹͳΓ·ͨ͠. • ʮ28ࡀͷ͕Maxʯʮ29ࡀ͔ΒԼ͕ͬͯΔʯͨΓ͕ϦΞϧ. ※ΞεϦʔτͷମత࠷ߴை26ʙ28ࡀͱݴΘΕ͍ͯ·͢ • ͱ͍͑28ࡀͷຊྥଧ্͕͕ͬͯΔͷ, ͳΜ͔ո͍͠.
͓ͦΒ͘୭͔ͷʹҾͬுΒΕ͍ͯΔ.
Γ͠ɾվળϙΠϯτ • ࠷ޙͷྨ, ϕΠζҎ֎ࢼ͍ͨ͠. • ʮ28ࡀΛʹਰ͑ΔʯϙδγϣϯʹΑͬͯҧ͏આ͋Δ. ͷͰʮ্ͷʯΛٻΊΔλεΫ͕͍͍͔͋ͬͯ. • 2020ͷϝδϟʔϦʔάྫͷͷࢼ߹ͳͷͰ, ༧ଌͦ͜ʹ߹Θ͍ͤͨʢ2ͰׂͬͯऴΘΔʁwʣ
• ͱ͍͏ͷ͕PyCon JP 2020·ͰʹͰ͖ͯΔͣʢVer. 25ʮTsurageʯͰʣ
ଓ͖PyCon JP 2020Ͱʂ #͓͠·͍ #͝ਗ਼ௌ͋Γ͕ͱ͏͍͟͝·ͨ͠ Shinichi Nakagawa(Twitter/Facebook/etc… @shinyorke)