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
プロスペクトをデータで紹介
Search
kuma127
March 27, 2019
Research
2
2.7k
プロスペクトをデータで紹介
3/27 BaseballPlayStudyのLTスライド
kuma127
March 27, 2019
Tweet
Share
More Decks by kuma127
See All by kuma127
ReactPyを使ってreact likeにUIをPythonで実装する
kuma127
0
24
私と#pyhack
kuma127
0
280
配信チームの準備と当日進行について
kuma127
0
360
石川雅規のことをもっと知ってもらいたいLT
kuma127
0
1.5k
Beginning PyData with Django and Jupyter
kuma127
3
5.7k
DjangoとJupyterで捗るPyData
kuma127
1
1.6k
趣味駆動学習のススメ〜野球の場合〜
kuma127
0
340
気軽に手に入る成績データで野球分析をしてみた
kuma127
0
2.1k
Python is simple and deep
kuma127
0
54
Other Decks in Research
See All in Research
論文紹介:Not All Tokens Are What You Need for Pretraining
kosuken
0
200
多言語カスタマーインタビューの“壁”を越える~PMと生成AIの共創~ 株式会社ジグザグ 松野 亘
watarumatsuno
0
130
EarthSynth: Generating Informative Earth Observation with Diffusion Models
satai
3
370
Pythonでジオを使い倒そう! 〜それとFOSS4G Hiroshima 2026のご紹介を少し〜
wata909
0
900
最適化と機械学習による問題解決
mickey_kubo
0
180
大学見本市2025 JSTさきがけ事業セミナー「顔の見えないセンシング技術:多様なセンサにもとづく個人情報に配慮した人物状態推定」
miso2024
0
160
2025/7/5 応用音響研究会招待講演@北海道大学
takuma_okamoto
1
220
Learning to (Learn at Test Time): RNNs with Expressive Hidden States
kurita
1
260
SegEarth-OV: Towards Training-Free Open-Vocabulary Segmentation for Remote Sensing Images
satai
3
310
Towards a More Efficient Reasoning LLM: AIMO2 Solution Summary and Introduction to Fast-Math Models
analokmaus
2
890
AIスパコン「さくらONE」のLLM学習ベンチマークによる性能評価 / SAKURAONE LLM Training Benchmarking
yuukit
2
680
Vision and LanguageからのEmbodied AIとAI for Science
yushiku
PRO
1
560
Featured
See All Featured
Save Time (by Creating Custom Rails Generators)
garrettdimon
PRO
32
1.6k
Six Lessons from altMBA
skipperchong
28
4k
Creating an realtime collaboration tool: Agile Flush - .NET Oxford
marcduiker
32
2.3k
Visualization
eitanlees
148
16k
The Cost Of JavaScript in 2023
addyosmani
53
9k
Fantastic passwords and where to find them - at NoRuKo
philnash
52
3.4k
Making the Leap to Tech Lead
cromwellryan
135
9.6k
Building Better People: How to give real-time feedback that sticks.
wjessup
368
20k
Become a Pro
speakerdeck
PRO
29
5.5k
ReactJS: Keep Simple. Everything can be a component!
pedronauck
667
120k
Responsive Adventures: Dirty Tricks From The Dark Corners of Front-End
smashingmag
252
21k
Why You Should Never Use an ORM
jnunemaker
PRO
59
9.6k
Transcript
2019/3/27 ࠓͷϒϨΠΫɺདྷͷޒྠද ୭ͩʁࠓͷϓϩεϖΫτΛσʔλͰհ ͘·ͽ TwitterIDɿ@kumappp27
ࣗݾհ ✤ ʮ͘·ͽʯͱ͍͍·͢ ✤ TwitterID:@kumappp27 ✤ ݹాPM͔࣌ΒͷͪͳϠΫ ✤ খହ߂खதଜ༔ฏัखͱಉ͍ͷΪϦฏੜ· Εʢฏ2ʣ
✤ ࣅ͍ͯΔͱݴΘΕ͍ͯͨळ٢ख͕Ҡ੶ͯ͠γϣοΫ
ຊͷ͓
ࠓͷϒϨΠΫɺདྷͷޒྠද ୭ͩʁࠓͷϓϩεϖΫτΛσʔλͰհ
͓ʹ͍ͭͯ ✤ ຖɺ͜ͷ࣌ظ͋Δ͋ΔͷҰͭͱͯ͠एखͷ׆༂Ͱ ϙδΔɺͱ͍͏ͷ͕͋Δͱࢥ͍·͢ ✤ ͋ͱɺͳΜ͔ΜͰଞٿஂͷબखͰए͍બख͕ϒ ϨΠΫ͢Δͷت͍͜͠ͱ…ͳؾ͕͠·͢ ✤ ͦ͜Ͱɺ͜ͷΛआΓͯࢲΠνΦγͷएखΛɺσʔλ Λ༻͍ͯհ͠Α͏ɺͱ͍͏LTͰ͢
հ͢Δʹ͋ͨͬͯ ✤ جຊతʹࢲͷಠஅͰϐοΫΞοϓ͠·͕ͨ͠ɺҰԠҎԼͷ ݅ͷݩϐοΫΞοϓ͠·ͨ͠ 1. ࠓͰ3·ͰʢܦྺΘͣʣ 2. ৽ਓԦࢿ͕֨͋Δ 3. υϥϑτ1Ґ͡Όͳ͍
✤ 1ͱ2͚ͩͩͱ༗໊ॴͷհʹͳͬͪΌ͏ͱࢥ͏ͷͰɺυϥϑ τ1Ґআ֎͠·ͨ͠
Special Thanks ✤ ࠓճσʔλΛ͓आΓͨ͠ఏڙݩͷαΠτ༷ͪ͜Βɹ ʮϓϩٿσʔλFreakʯ༷ ✤ ϦϯΫɿhttps://baseball-data.com/ ✤ 1܉ͷσʔλͪΖΜ2܉ͷσʔλॆ࣮͍ͯ͠Δૉ Β͍͠αΠτͰ͢
Ͱͬͦ͘͞1ਓ
ౡ౦༸Χʔϓɾࡔকޗ 19985݄29ੜ·Εʢ20ࡀʣ 2016υϥϑτ4Ґ ӈࠨଧɹัख
ࡔબख͜Μͳਓ ✤ ัखͳ͕Βߴ͍ଧܸηϯεɺ 50m࠷6ඵ3Λه͢Δ΄Ͳ ͷढ़Λซͤͭ ✤ ϧʔΩʔΠϠʔͷ2017ʹެ ࣜઓͰώοτΛଧ͕ͬͨɺ͜ Εౡͷߴଔ৽ਓัखͱ͠ ͯҥּ༤ࢯҎདྷͷҒۀ
ʢWikipediaΑΓൈਮʣ
ͦΜͳࡔબखͷηʔϧεϙΠϯτ…
ัखͰ͋Γͳ͕Βͷߴ͍ଧܸྗ ✤ 2ؒͷ2܉ͪ͜Β ଧ ຊྥଧ ग़ྥ ଧ OPS 2017 .298
1 .359 .400 .759 2018 .329 4 .372 .547 .919
ผࢦඪ͔Βࡔબखͷੌ͞ΛଌΔ ✤ ଧऀͷ߈ܸྗΛଌΔࢦඪͱͯ͠wOBA※ͱ͍͏ͷ͕͋ΔͷͰɺͦͪΒΛ༻͍ ͯࢉग़ͨ͠σʔλ͕ͪ͜Β ✤ ͜ͷࣈɺߴ͍΄Ͳྑ͘ɺࣈͷεέʔϧ͕ग़ྥͱಉ͡Α͏ʹग़དྷ͍ͯΔ ※wOBAͱ 1ଧ੮ͨΓͷಘग़ྗΛଌΔࢦඪɻ ҆ଧ࢛ٿͳͲͷग़ྥ݁ՌʹಠࣗͷॏΈ͚ͮΛͯ͠ࢉग़͢Δɻ ʢWikipediaΑΓɺࠓճσʔλͷ্ࣦؔࡦग़ྥͷهແࢹ͠·͢ʣ
wOBA 2017 .336218 2018 .397907
…Θ͔ΓͮΒ͍ͷͰྺͷऀୡͱൺֱͯ͠Έ·͠ΐ͏ ࠓΛͱ͖Ί͘൴Βʹ͝ొئ͍·ͨ͠ ※൴Βօଧ੮͕όϥόϥͰ͕͢ɺ wOBA1ଧ੮͋ͨΓͷࢦඪͳͷͰެฏ͋͞ΔఔอͯΔ͔ͱ ಡചδϟΠΞϯπ Ԭຊਅ બख ౡ౦༸Χʔϓ ླ બख
౦ژϠΫϧτεϫϩʔζ ࢁాਓ બख
൴Βͷߴଔ1,2ͱൺֱͯ͠ΈΔ ※ิɹླ2ʹ 1܉Ͱ68ଧ੮Ͱଧ.344Λه͍ͯ͠Δ બख໊ wOBA ࡔ 1 .336 2
.398 Ԭຊ 1 .293 2 .360 ླ 1 .304 2 .323 ࢁా 1 .295 2 .335
ૉΒ͍͠ ✤ ྺͷऀͱൺͯḮ৭ͳ͍ʢϦʔάҧ͏ ͷͰɺࢀߟఔͰ͕͢ʣ ✤ ͦΕͲ͜Ζ͔ɺ্ճͬͨΛ͍ͯ͠Δ ✤ ઌ΄Ͳड़͕ͨɺ͜ΕͰ͍ͯัख͔ͩΒڪΕଟ͍
ͦΜͳࡔબखʹΈࣄ͕ ✤ ౡͷัख͕ް͗͢Δ ✤ ࠓग़ػձΛٻΊͯ֎ʹઓ͍ͯ͠Δ༷ ϕετφΠϯ ܦݧ๛ͳϕςϥϯ ༗ג
ࡔબखͷظ ✤ όοςΟϯάͷ͍͍ัखͱͯ͠ɺΏ͘Ώ͘ͷຊද ೖΓΛظ ✤ ઌड़ͷྺͷऀօߴଔ4ʹେϒϨΠΫΛՌͨ ͨ͠ͷͰͦͷྲྀΕʹΔͳΒϒϨΠΫདྷʁ
࣍ʹ2ਓ
౦ژϠΫϧτεϫϩʔζɾԘݟ ହོ 19936݄12ੜ·Εʢ25ࡀʣ 2017υϥϑτ4Ґ ӈӈଧɹ֎ख
Ԙݟબख͜Μͳਓ ✤ ग़͕૬ߴֶߍͰɺ ग़ͷޙഐ ✤ #ϠόΠΑԘݟ ✤ ࡢͷΞδΞΟϯλʔ ϦʔάͰ࠷༏लଧऀ Λड
ͦΜͳԘݟબखͷηʔϧεϙΠϯτ…
ɾ߈ڞʹѹతͳ࣮ ✤ ϧʔΩʔΠϠʔͷࡢɺ1܉Ͱଧ.040ͱۤ͠Μͩ ͕ɺ2܉ࠓقͷΦʔϓϯઓͰѹతͳΛه ଧ HR ౪ྥ OPS BB/K ’18
Πʔελϯ .329 9 22 1.011 0.51 ’18 Οϯλʔ Ϧʔά .392 4 5 1.209 1.17 ’19 Φʔϓϯઓ .385 2 12 1.025 0.44
ࠓճྺͷऀͱൺֱ ✤ ࡢେ׆༂ͨ͠େଔɾࣾձਓग़ͷબखͷ1ͷ2܉ͱൺͯΈΔ ʢ˞ϝϯπ͕ͨΒΰπ͍Ͱ͕͢ɺؾͷ͍ͤʣ ࡛ۄϥΠΦϯζ ࢁึߴ બख ઍ༿ϩοςϚϦʔϯζ Ҫ্࠸ બख
ԣDeNAϕΠελʔζ ٶ㟒හ બख
൴Βͷ1ͷ̎܉ͱൺͯΈΔ બख໊ wOBA Ԙݟ .4294 ࢁ .4332 Ҫ্ .4490 ٶ࡚
.3863
ࢁʹΞδϟίϯάʹுΓ߹͍ͬͯΔ ✤ λΠϓతʹ3໊ͱԘݟબखͱͰશવҧ͏ͷ͕ͩɺ ͦΜͳ3ਓͷதʹׂͬͯೖΔΑ͏ͳ߈ܸྗΛඋ͍͑ͯ Δ ✤ #ϠόΠ ✤ ͜ͷ3ਓ͍ͣΕ3ʙ5ลΓʹϒϨΠΫ͍ͯ͠Δ ͷͰɺԘݟબखࠓޙͷϒϨΠΫ͕ظͰ͖Δ͔͠
Εͳ͍
ྫʹΑͬͯԘݟબखʹΈࣄ͕ ✤ ϠΫϧτͷ֎ϨΪϡϥʔਞ͕൫ੴ͗͢Δ HRهอ࣋ऀ ϝδϟʔؼΓ 3ׂόολʔ ✤ ·ͣπʔϓϥτϯى༻͔Βͷελʔτ͔…ʁ 3ׂόολʔ ϑΝʔετकΕΔ
Ԙݟબखͷظ ✤ νʔϜঢ়گతʹଈϨΪϡϥʔݫ͍͕͠ɺߴྸԽ͕ਐ ΜͰ͍ΔϠΫϧτ֎ਞͷएฦΓͷτοϓʹͳΕΔΑ ͏ͳ׆༂Λظ ✤ ૉΒ͍͠߈ܸྗʹΦʔϓϯઓ౪ྥԦͷ࣮ྗΛ׆͔ͤ Εɺ͍ۙࢁాબखͱͷτϦϓϧεϦʔίϯϏ͕ੜ ·ΕΔ…͔͠Εͳ͍
·ͩ·͍ͩΔͧʂ
ٽ͖ͷՃհ ✤ հ͔͚ͨͬͨ͠Ͳ࣌ؒͷ߹্Β͟ΔΛಘͳ͔ͬ ͨબखΛ໊લͱɺηʔϧϙΠϯτΛҰݴड़͍͖ͯ· ͢
ࡕਆλΠΨʔεɾਅख ✤ ࠷150km/hͷٿʹεϥΠμʔɺ ΧʔϒɺνΣϯδΞοϓΛૢΔظ ͷएखख ✤ ηʔϧϙΠϯτʮ੍ٿྗʯ ✤ ࡢ2܉Ͱ27ΠχϯάΛ࢛͛ٿ Θ͔ͣʮ3ʯ
✤ ͜ΕࡢେϒϨΠΫΛՌͨͨ͠ɺ ΦϦοΫεࢁຊ༝৳खͷ1ͷ ʹඖఢ͢Δ 19985݄25ੜ·Εʢ20ࡀʣ 2016υϥϑτ4Ґ ӈӈଧɹख
ઍ༿ϩοςϚϦʔϯζɾӬক࢘ख ✤ ࠨ͔Β࠷150km/hΛ܁Γग़͢ ٿϦϦʔόʔ ✤ ηʔϧεϙΠϯτɺʮୣࡾৼྗʯ ✤ 2܉Ͱ22ࢼ߹ʹ͛ ୣࡾৼɿ12.43 K/BBɿ7.25
ɹͷѹרϐονϯάΛ൸࿐ ✤ ࠓɺʮڪාʯΛެදɻݸ ਓతʹԠԉ͍ͨ͠બखɻ 19933݄2ੜ·Εʢ26ࡀʣ 2017υϥϑτ6Ґ ࠨࠨଧɹख
Ҏ্Ͱ͢ ✤ օ͞Μɺࣗਪ͠ͷएखબख͍·ͨ͠Ͱ͠ΐ͏͔ʁ ࠓճڈͷ2܉ΛݩʹϐοΫΞοϓͨ͠ͷͰɺ ࣈʹݱΕͯͳ͍͚ͲظͰ͖Δएखબख·ͩ·ͩ ͍Δͱࢥ͍·͢ ✤ ʮ͜ͷબख͍͍ͧʂʯͬͯબख͕͍Εͥͻͥͻڭ͑ ͖͍ͯͨͰ͢ʙ
͝੩ௌɺ͋Γ͕ͱ͏͍͟͝·ͨ͠ʂ