ãã®è¨äºã¯Classi developers Advent Calendar 2021ã®18æ¥ç®ã®è¨äºã§ãã
æ¨æ¥ã¯åºç¤ã¤ã³ãã©ãã¼ã ã®ããããã«ãããããã¯ã¨ã³ãã¨ã³ã¸ãã¢ãåºç¤ã¤ã³ãã©ãã¼ã ã«ç°åãã¦åå¹´ã»ã©çµã£ã話ãã§ããã
ããã«ã¡ã¯ããã¼ã¿AIé¨ã§ãã¼ã¿ãµã¤ã¨ã³ãã£ã¹ãããã¦ããé«æ¨ã§ãã
å¼ç¤¾ã§ã¯é¡§å®¢ã§ããå
çãçå¾ãä¿è·è
ããClassiã®æ©è½ãå¥ç´ã«é¢ããåãåãããæ¥ã
é ãã¦ããã¾ãã
ãããã®åãåããã®å
容ãåæããClassiã®ç¾ç¶ã®èª²é¡ãä»å¾è§£æ±ºãã¦ããããã®æ½çãªã©ã社å
ã§æ¤è¨ãã¦ãã¾ãã
ä»åã¯åãåããå 容ãè¨èªå¦çæè¡ã®ä¸ã¤ã§ãããããã¯ã¢ãã«ã使ã£ã¦åæããå 容ã«ã¤ãã¦ãç´¹ä»ãã¾ãã
ãªãåæããå¿ è¦ããã£ãã®ãï¼
Classiã¸ã®åãåããããã®å¯¾å¿ã®å
容ã¯ãæ
å½è
ã«ãã£ã¦ããã¹ãåãããç¶æ
ã§ç®¡çããã¦ãã¾ãã
å¼ç¤¾ã®ã«ã¹ã¿ãã¼ãµãã¼ãã»ã«ã¹ã¿ãã¼ãµã¯ã»ã¹ãã¼ã ã¯ããããã®å
容ãåºã«æ¬¡å¹´åº¦ã®æ½çãæ¤è¨ããããå®æ½ããæ½çãã©ãã ã£ããã¨ããå¹æãåæãã¦ãã¾ãã
年度ãå¤ãã3æã4æã§ã®åãåãããå¢å å¾åã«ããããã®ä¸ã§ãå¦æ ¡å
¨ä½ã®å©ç¨ã«é¢ããè¨å®ã»ç»é²ã®åãåãããæãå¤ããªã£ã¦ãã¾ãã
æ¯å¹´ããããã®åãåãããç®è¦ã§ãã§ãã¯ã»åé¡ããæ½çãæ¤è¨ãã¦ããç¶æ
ã ã£ãã®ã§ãããä½æ¥ã«ã¯è¨å¤§ãªæéã¨å´åãããã£ã¦ãã¾ãã¾ãã
ããã§ãããç¨åº¦èªåçã«åæã»åé¡ã§ããªããã¨ããå¿ è¦æ§ãåºã¦ãã¦ãããä»åãããã¯ã¢ãã«ã使ã£ãåæã試ã¿ã¾ããã
ãããã¯ã¢ãã«ã¨ã¯ï¼ãªã使ã£ãã®ãï¼
ãããã¯ã¢ãã«ã¯ãææ¸éåããããã«å«ã¾ããåèªã®å ±èµ·æ§ãããã©ããªãããã¯ï¼è©±é¡ãåéãªã©ã大ä½ã®ãæå³ãã®ãããªãã®ï¼ãåå¨ããã®ããåææ¸ã¯ã©ã®ãããã¯ã«ã¤ãã¦ã®ãã®ãããèªåã§æ¨å®ãããã¨ãã§ãã¾ãã
ããã¾ã§æ§ã
ãªææ³ãææ¡ããã¦ãããææ¸ã®æ¤ç´¢ãåé¡ãã¾ããé³æ¥½ã®æè©æ¢ç´¢*1ãã²ã¼ã ã®ãããã¢ã¼ãã¿ã¤ãã®æ½åº*2ãªã©ãæ§ã
ãªåéã¸å¿ç¨ããã¦ãã¾ãã
ç§ããã¹ãåé¡ã«å¿ç¨ããç 究ããã¦ãã¾ã*3ã
以ä¸ã¯ä»£è¡¨çãªãããã¯ã¢ãã«ã§ãã
ãããã¯ã¢ãã« | ç¹å¾´ | åèæç® | åèã³ã¼ã |
---|---|---|---|
LSI: Latent Semantic Indexing | ã»ç¹ç°å¤å解ã«ããæç« ã®å§ç¸® ã»1ææ¸ã«1ããã㯠|
Deerwester et al (1990)*4 | Gensim |
PLSI: Probabilistic Latent Semantic Indexing | ã»LSIã®ç¢ºçã¢ãã«å ã»1ææ¸ã«è¤æ°ããã㯠|
Hofmann (1999)*5 | PyPI |
LDA: Latent Dirichlet Allocation | ã»PLSIã®ãã¤ãºå ã»æ°è¦ææ¸ã®ãããã¯æ¨å®ãå¯è½ |
Blei et al (2003)*6 | Gensim |
HDP: Hierechical Dirichlet Process | ã»LDAã®ãã³ãã©ã¡ããªãã¯å ã»ãããã¯æ°ãèªå決å®å¯è½ |
Teh et al (2006)*7 | Gensim |
DTM: Dynamic Topic Model | ã»æç³»åãã¼ã¿ã¸ã®é©ç¨ | Blei et al (2006)*8 | Gensim |
BTM: Biterm Topic Model | ã»çãæç« ã¸ã®é©ç¨ | Yan et al (2013)*9 | GitHub |
ä¸æ¹ã§ãåãåããã®å
容ã«ã¯Classiã®æ©è½ãå¥ç´ã«é¢ããè¤æ°ã®è³ªåãåæã«å«ã¾ãããã¨ãããã¾ãã
ããã¦ããã®å
容ã¯ææã«ãã£ã¦ãå¤åãã¾ãã
ããã§ä»åã¯ãï¼1ï¼èªåã§æ軽ã«å 容ãææ¡ã§ãããï¼2ï¼åä¸ã®åãåããããè¤æ°ã®ãããã¯ãæ½åºãããã¨ãã§ãããï¼3ï¼å 容ã®æç³»åå¤åã追ããã¨ãã§ããã ã¨ããç¹ãèæ ®ãDTMï¼Dynamic Topic Modelï¼ã使ç¨ãã¾ããã
ä¸å³ã¯DTMã«ãããããã¯ã®æ¨å®ã¨æç³»åå¤åã®ä¾ã§ãã
æå»tã«ããããããææ¸éåã®ãããã¯ã«åºç¾ããåèªï¼æ¨ªè»¸ï¼ã¨ããããã®åèªããããã¯ã«åºç¾ãã確çï¼ç¸¦è»¸ï¼ã表ãã¦ãã¾ãã
ã¾ããåææ¸ã§ããããã®ãããã¯ãåºç¾ãã確çãæ¨å®ããã¾ãã
DTMã«ããåãåããå 容ã®åæ
æ¦è¦
2020å¹´ã¨2021å¹´ã®3æã4æã«å¯ããããè¨å®ã»ç»é²ã«ã¤ãã¦ã®åãåããããããã¯åæããå
容ã®å¤åãèå¯ãã¾ããã
DTMã使ã£ããããã¯åæã®æé ã¨ç®çã¯ä»¥ä¸ã«ãªãã¾ãã
ãªãããã¹ã¦ã®åæã¯Pythonã§è¡ãã¾ããã
- DTMã«ãããããã¯ã®æ¨å®
- 2020å¹´ã¨2021å¹´ã®ããããã®åãåããã®ãããã¯ã¨ãã®æç³»åå¤åãæ¨å®ãããã
- ãããã¯ã¸ã®ã©ãã«ä»ã
- ãããã¯ã®å 容ãææ¡ãããããããã
- ãããã¯ãã¨ã®åãåãã件æ°ãéè¨
- ã©ã®ãããã¯ã®åãåãããå¤ããææ¡ãããã
- 2020å¹´ã¨2021å¹´ã®ãããã¯éé¡ä¼¼åº¦ã®è¨ç®
- é¡ä¼¼åº¦ã®ä½ããããã¯ï¼åå¹´ã®ç¹æãªãããã¯ï¼ãæ½åºãããã
- ããã·ã¥ãã¼ãã®ä½æ
- æé 1ã4ã®çµæãå¯è¦åãããã
åææé ã®è©³ç´°
1. DTMã«ãããããã¯ã®æ¨å®
DTMã«ãããããã¯ã®æ¨å®ã¯ãGensimã®models.ldaseqmodelã¨ããã©ã¤ãã©ãªã使ç¨ãã¾ããã
ãããã¯ã®æ¨å®ã§ä½¿ç¨ããã¡ã½ããã¨å
¥åãã©ã¡ã¼ã¿ã¯ä»¥ä¸ã«ãªãã¾ãã
dtm_model = LdaSeqModel(corpus, id2word, time_slice, num_topics)
- corpusï¼ã³ã¼ãã¹
- ææ¸æ¯ã®åèªIDã¨ãã®åºç¾åæ°ãæã¤ã¿ãã«ãªã¹ã
- id2wordï¼è¾æ¸
- åèªã¨ãããä¸æã«èå¥ããIDã®è¾æ¸
- time_sliceï¼ã¿ã¤ã ã¹ã©ã¤ã¹
- æå®ãããæéããã¨ã®ææ¸æ°ã®ãªã¹ã
- num_topicsï¼ãããã¯æ°
- ãããã¯ã®æ°ã¯å¯å¤
è¾æ¸ãã³ã¼ãã¹ã®è©³ç´°ã«ã¤ãã¦ã¯çç¥ãã¾ãããããããä½æããéã«éè¦ã¨ãªãåèªã®æ½åºå¦çã«ã¤ãã¦å¾è¿°ãã¾ãã
ã¿ã¤ã ã¹ã©ã¤ã¹ã¨ã¯ããããã¯ã®å¤åãæå®ãããæéãã§åå²ããããã®ãã®ã§ãä»åã¯ãã®ãæéããæææ¥ãæ¥ææ¥ã®1é±éãã¨ã«ãã¾ããã
é±æ¯ã®åãåãã件æ°ã以ä¸ã®å ´åãtime_slice = [3, 5, 1]
ã¨ãªãã¾ãã
- 2021/03/01ï¼æï¼ã03/07ï¼æ¥ï¼ï¼3件
- 2021/03/08ï¼æï¼ã03/14ï¼æ¥ï¼ï¼5件
- 2021/03/15ï¼æï¼ã03/21ï¼æ¥ï¼ï¼1件
ããã«ããããã¯æ°ã¯åæãåé¡ãããç²åº¦ã«ãã£ã¦æåã§è¨å®ããå¿
è¦ãããã¾ãã
ä»åã¯ä»¥ä¸ã®çç±ã§ä¸¡å¹´ã¨ããããã¯æ°ã10ã¨ãã¾ããã
- 2020å¹´ã®åãåããã調æ»ããéãç´10åã®éè¦ãªã«ãã´ãªã«åé¡ããã
- å¤ãããã¨ãããã¯ã®æå³ãç´°ããåæ£ããã§ãã¯ããã®ã大å¤
åèªã®æ½åºå¦çã«ã¤ãã¦
ãããã¯ã¢ãã«ã§ã¯ãããã¯ã®æ¨å®ç²¾åº¦ãåä¸ãããããã«ãé¢é£ã®ããåèªã®å
±èµ·æ§ãé«ãããã¨ãéè¦ã«ãªã£ã¦ãã¾ãã
ä¸æ¹ã§ãæ¥æ¬èªã«ã¯æ¬¡ã®ãããªæ§è³ªãããã¾ãã
ååè©Nã対象åéã®éè¦ãªæ¦å¿µã表ãã¦ãããªããæ¸ãæã¯Nãé »ç¹ã«åç¬ã§ä½¿ãã®ã¿ãªãããæ°è¦ãªæ¦å¿µã表ã表ç¾ã¨ãã¦Nãå«ãè¤ååè©ãä½ãã ããã¨ãå¤ãã*10
ãã®ãååè©Nãã¨ãååè©Nãå«ãè¤ååè©ãã¯é¢é£ããæ¦å¿µã表ãã¦ãããã¨ãå¤ãã§ãã
ä¾ãã°ãClassiã®ã³ãã¥ãã±ã¼ã·ã§ã³æ©è½ã§ãããæ ¡å ã°ã«ã¼ããã§ã¯ããã°ã«ã¼ããã¨ããåèªãä¸å¿ã«ããèªåä½æã°ã«ã¼ããããä»»æä½æã°ã«ã¼ãããªã©ã®æ©è½ãåå¨ããåä¸ã®åãåããå ã§åºç¾ãããã¨ãå¤ãã§ãã
ãã®ä¾ã®å ´åããååè©Nãã¨ãååè©Nãå«ãè¤ååè©ãã¯ä»¥ä¸ã«ãªãã¾ãã
- ååè©ï¼æ ¡å ãã°ã«ã¼ããèªåãä½æãä»»æ
- è¤ååè©ï¼æ ¡å ã°ã«ã¼ããèªåä½æã°ã«ã¼ããä»»æä½æã°ã«ã¼ã
ãããã®åèªã¯ãã¹ã¦é¢é£ãã¦ãããåä¸ã®ãããã¯ã«åé¡ããããã¨ãæã¾ããããããã®ãããªååè©ã¨è¤ååè©ããã¹ã¦æ½åºããããã«ãã¾ããã
æ½åºå¦çã«ã¯ãå½¢æ
ç´ è§£æå¨ã®MeCabã¨å½¢æ
ç´ è§£æå¨ã®çµæãåºã«è¤ååè©ãæ½åºãããã¨ãå¯è½ãªTermExtractã使ç¨ãã¦ãã¾ãã
2. DTMã«ãããããã¯ã®æ¨å®
æé 1ã§ä½æãããdtm_model
ã«ãããç¹å®ã®æéã®ãããã¯ã«åºç¾ããåèªã¨åºç¾ç¢ºçããªã¹ãã§åºåãããã¨ãã§ãã¾ãã
ãã®ã¡ã½ããã¨å
¥åãã©ã¡ã¼ã¿ã¯ä»¥ä¸ã«ãªãã¾ãã
dtm_model.print_topic(topic, time, top_terms)
- topic
- ãããã¯ã®IDï¼ãããã¯æ°10ã®å ´å0~9ï¼
- time
- ã¿ã¤ã ã¹ã©ã¤ã¹ã§è¨å®ãããæéãã®é çª
- top_terms
- åºåããåèªã®æ°
ãã®ã¡ã½ããã«ããåºåçµæãç¨ãã¦ããããã¯ãã¯ã¼ãã¯ã©ã¦ãã§å¯è¦åãã¾ããã
ããã¦ãæ¯è¼ç大ãã表ãããåèªãåºã«ã©ãã«ãä»ä¸ãã¾ããã
æåã®é±ï¼time=0ï¼ã®ãããã¯ã®ã¯ã¼ãã¯ã©ã¦ãã¯ä»¥ä¸ã«ãªãã¾ãã
3. ãããã¯ãã¨ã®åãåãã件æ°éè¨
dtm_model
ã§ã¯ãåãåãããã¨ã«ãããã¯ã®åºç¾ç¢ºçãåºåãããã¨ãã§ãã¾ãã
ãã®ã¡ã½ããã¨å
¥åãã©ã¡ã¼ã¿ã¯ä»¥ä¸ã«ãªãã¾ãã
dtm_model.doc_topics(doc_number)
- doc_number
- ã³ã¼ãã¹ãè¾æ¸ãä½æããéã«å ¥åããææ¸ã®é çª
åãåãããã¨ã«æãåºç¾ç¢ºçã®é«ããããã¯ããã®åãåããã«ããããããã¯ã¨ãããããã¯ãã¨ã®åãåãã件æ°ãéè¨ãã¾ããã
以ä¸ã®è¡¨ã¯ä»¶æ°ã®å¤ãé ã«ãããã¯ã並ã¹ãçµæã«ãªãã¾ãã
2020å¹´3æã4æ
ãããã¯No | ã©ãã« |
---|---|
2 | 年度æ´æ°ã®åã¹ãããä½æ¥ |
10 | çå¾æ å ±ã®ç»é² |
1 | çå¾ãå çã®æ å ±ç»é², Excelãã¡ã¤ã«ã®UPL/DL |
6 | IDã¨PW, ãã°ã¤ã³, æå¾ ã³ã¼ã, ç¡åæä¾ |
4 | 管ç責任è , 管çè ã®æ¨©é, çå¾ã«ã«ã |
3 | æ ¡å ã°ã«ã¼ãã®è¨å®ãé ä¿¡, ææ¥ã®ç»é² |
7 | å¦ç¿åç», å¦ç¿è¨é², ããã¼çå¾ |
9 | ç¶ç¶, å©ç¨åæ¢, ä¿è·è |
5 | èå¥çªå·, çå¹´ãé²å¦ããçå¾ã®ç»é², 人æ°ç¢ºå® |
8 | 模試ãã¼ã¿é£æº |
2021å¹´3æã4æ
ãããã¯No | ã©ãã« |
---|---|
2 | çå¾æ å ±ã®ç»é², Excelãã¡ã¤ã«ã®UPL/DL |
3 | 年度æ´æ° |
5 | 年度æ´æ° |
4 | IDã¨PW, ãã°ã¤ã³, æå¾ ã³ã¼ã |
8 | 模試ãã¼ã¿é£æº |
7 | å©ç¨åæ¢, å¦ç¿åç»ãã㯠|
6 | æ ¡å ã°ã«ã¼ã, ã¢ã³ã±ã¼ã |
1 | 管ç責任è , å¦ç¿åç», å¦ç¿è¨é², é²è¦§æ¨©é |
10 | ææ¥ç»é², å çç»é², 帳票ç»é² |
9 | 人æ°ç¢ºå® |
4. ãããã¯éé¡ä¼¼åº¦ã®è¨ç®
æé 2ã§ä½¿ç¨ããdtm_model.print_topic
ãç¨ãã¦ããããã¯ã«å«ã¾ããåèªã¨ãã®åºç¾ç¢ºçã50ååºåãã2020å¹´ã¨2021å¹´ã®ãããã¯éã®é¡ä¼¼åº¦ãã³ãµã¤ã³é¡ä¼¼åº¦ã§è¨ç®ãã¾ããã
ä¸å³ã¯ãããã¯éã®é¡ä¼¼åº¦ããã¼ããããã«ãã£ã¦å¯è¦åããçµæã§ãã
è¡ã2020å¹´ãåã2021å¹´ã®ãããã¯ã示ãã¦ãããããããã®é¡ä¼¼åº¦ã®å¤ã¨å¤ã大ããã»ã©è²ãæ¿ã表示ããã¦ãã¾ãã
ä¾ãã°ã2021å¹´ã®topic9ã¯åå¹´ã®ä»ãããã¯ã«æ¯ã¹2020å¹´ã®ãããã¯ã¨ã®é¡ä¼¼åº¦ããã¹ã¦0.5æªæºã¨ä½ããªã£ã¦ãã¾ãã
ãã®ãã¨ãããtopic9ã¯2021å¹´ç¹æãªãããã¯ã ã¨ãããã¨ãåããã¾ãã
5. ããã·ã¥ãã¼ãã«ããå¯è¦å
æé 1ã4ã®çµæãTableauã®ããã·ã¥ãã¼ãã§å¯è¦åãã¾ããã
ä¸å³ã¯2021å¹´ã®ããã·ã¥ãã¼ãã§ãã
ãã®ããã·ã¥ãã¼ãã§åãããã¨ã¯ä»¥ä¸ã«ãªãã¾ãã
- å年度ã®ãããã¯ã表ãåèªããã®åºç¾ç¢ºçï¼å³ä¸â ï¼
- ãããã¯ãã¨ã®åèªã®é±æ¬¡å¤åï¼å³ä¸â¡ï¼
- åãããã¯ã«è©²å½ããåãåããå 容ï¼å³ä¸â¢ï¼
- ãããã¯ãã¨ã®åãåãã件æ°ï¼å³ä¸â£ï¼
- ãããã¯éã®é¡ä¼¼åº¦ï¼æé 4ã®ãã¼ãããããå¥ãã¼ã¸ã§ä½æï¼
å 容å¤åã®èå¯
以ä¸ã®åæçµæãåºã«ã2020å¹´ã¨2021å¹´ã®å
容ã®å¤åãèå¯ãã¾ããã
ãã¼ã¸ã®é½åä¸è©³ç´°ã¯çãã¾ãããããã¤ãèå³æ·±ãçµæã以ä¸ã§ç´¹ä»ãã¾ãã
ã年度æ´æ°ãã«ãããåã¹ãããã®åãåããã®æ¸å°
æé 3ã®åãåãã件æ°ã®å¤ããããã¯ãè¦ãã¨ã両年ã¨ãã«æ¬¡å¹´åº¦ã®ã¦ã¼ã¶ã¼æ å ±ãæ´æ°ã»ç»é²ããã年度æ´æ°ãããæ å ±ç»é²ãã«ã¤ãã¦ã®ãããã¯ãä¸ä½3ä½ã¾ã§ã«å ¥ã£ã¦ãã¾ãã
2020å¹´ã¯ãããã®ãããã¯ã§ãã¹ããããããSTEPãã¨ããåèªãå«ãåãåãããå¤ãã£ãã®ã§ããã2021å¹´ã§ã¯ãããã®åèªãå«ãåãåãããæ¸å°ãã¦ãã¾ããã
ãã®çç±ã¨ãã¦ã¯ã2020å¹´ãã年度æ´æ°ã®æé ãã¹ããããã¨ã«èª¬æããã¬ã¤ããä½æããããããã¹ããããã¨ã«åãåããããã人ãå¢ããã¨èãããã¾ãã
ããã¦ã2021å¹´ã¯ãããã®ã¬ã¤ããåºã«ããä½æ¥ãåãã¦è¡ãå çã®æ°ãæ¸å°ããï¼ä½æ¥ãç解ãã¦ããå çãå¢å ããï¼ããããããã®åèªãå«ãåãåãããæ¸å°ããã¨èãããã¾ãã
ä¸å³ã¯2020å¹´3æã4æã®ã年度æ´æ°ãã«é¢ãããããã¯ã®æ¨ç§»ã§ãã
ãå©ç¨çå¾äººæ°ã®ç»é²ç¢ºå®ãã«ã¤ãã¦ã®åãåããã®å¢å
æé 4ã§è¿°ã¹ãããã«ã2021å¹´ã®topic9ã¯åå¹´ã®ä»ãããã¯ã«æ¯ã¹2020å¹´ã®ãããã¯ã¨ã®é¡ä¼¼åº¦ãä½ããªã£ã¦ããã2021å¹´ç¹æãªãããã¯ã¨ãªã£ã¦ãã¾ãã
ãã®ãããã¯ã¯å©ç¨çå¾äººæ°ã®ç»é²ç¢ºå®ã«é¢ãããã®ã§ããã®ä½æ¥æ¹æ³ãç· åã«ã¤ãã¦ã®åãåãããå¤ãã£ãã§ãã
ãã®çç±ã¨ãã¦ã¯ã2021å¹´ã¯4æã«å
¥ã£ã¦ãããããã®ç¢ºå®ä½æ¥ãå®äºããããåå¦æ ¡ã¸ã®é£çµ¡ãå¢ãã¦ãã¾ããã
ãã®çµæããããã®ç¢ºå®ä½æ¥ã«é¢é£ããåãåãããå¢ããã¨èãããã¾ãã
ä¸å³ã¯2021å¹´3æã4æã®å©ç¨çå¾äººæ°ã®ç»é²ç¢ºå®ã«é¢ãããããã¯ã®æ¨ç§»ã«ãªãã¾ãã
4æ以éã§ãã®ãããã¯ã«é¢é£ããåèªã®åºç¾ç¢ºçãä¸æå¾åã«ãªã£ã¦ããã®ãåããã¾ãã
æå¾ã«
ä»åã®åæçµæãããç¾ç¶ã®Classiã®è¨å®ã»ç»é²ã«é¢ãã課é¡ãæ½çã®å¹æãªã©ãèå¯ãããã¨ãã§ãã¾ããã
ãã®çµæããã¼ã ãè¶
ãã¦è¦ããã¨ã«ãã£ã¦ãä»å¹´ã®æ½çã¯ã©ãã ã£ãã®ããä»å¾ã©ãããæ½çãæã£ã¦ããã¹ãããèãããã¨ãã§ãã¾ãã
ä¾ãã°ãã«ã¹ã¿ãã¼ãµãã¼ããã¼ã ã§ã¯ä»åã®åæã§ãçå¾äººæ°ã®ç»é²ç¢ºå®ã年度æ´æ°ã«ããã¦åè¨å®ã®ææ¥ãåããã¥ãããã¨ãã年度æ´æ°ã®STEP3ã®åãåãããå¤ããã¨ãæããã«ãªã£ãããããããã®åãåããã10%æ¸ãããã¨ãç®æ¨ã«æ¹åæ´»åãé²ãã¦ãã¾ãã
ã¾ããPMMï¼ãããã¯ããã¼ã±ãã£ã³ã°ããã¼ã¸ã£ã¼ï¼ãã¼ã ã§ã¯é¡§å®¢èª²é¡ããããã¯ãã¸åæ ããããã«ãä»åã®åææé ãã¦ã¼ã¶ã¼ããå¯ããããæ©è½æ¹åã«ã¤ãã¦ã®è¦æã«å¿ç¨ãããã®å 容ã®æ·±å ããè¦æãå®ç¾ããããã®éçºã®åªå é ä½ä»ããæ¤è¨ãã¦ãã¾ãã
ãã®ããã«ãä»åã®åãçµã¿ãåãã¼ã ã§é²ãããã¦ããæ½çã«å°ãã§ãå¯ä¸ã§ãã¦ãããã¨ã¯ããã¼ã¿AIé¨ã¨ãã¦ãå¬ããéãã§ãã
åæçµæãæ¯åçæ¯ã«èãã¦ä¸ãããæè¦ä¸ããåãã¼ã ã®çæ§ã«ã¯ã¨ã¦ãæè¬ãã¦ãã¾ãã
ç¾å¨ã10æã¾ã§ã®åãåããå 容ã®åæãå®äºãã¦ãããä»å¾ã¯æ¥å¹´ã®2æã¾ã§ã®åãåããå 容ã®åæãå®æ½ããå¹´éã®åãåããã®å¾åãæ´çãã¦ããäºå®ã§ãã
ããã«ãæ¥å¹´åº¦ä»¥éã®åãåããå 容ã®ç¶ç¶çãªåæãããããã®çµæãåãã¼ã ã¸ç´ æ©ãéå ã§ããä»çµã¿ãæ§ç¯ãã¦ããããã¨èãã¦ãã¾ãã
ææ¥ã®Classi developers Advent Calendar 2021ã®æ
å½ã¯æ¨ªç°ããã§ãã
ãããããé¡ããã¾ãã
*1:Lyric Jumper: https://lyric-jumper.petitlyrics.com
*2:ãé転ãªã»ãã㢠ãã«ããããæ©æ¢°å¦ç¿ã¢ãã«ãç¨ãããããã®ã¢ã¼ãã¿ã¤ãæ½åºã¨ã²ã¼ã éç¨ã¸ã®æ´»ç¨: https://www.slideshare.net/RyoAdachi/deck-archetype-extraction-cedec2019
*3:é«æ¨è¼å½¦ï¼é«æ¨æ£åï¼å 使河åå¯æµ·ï¼ç°ä¸å¥æ¬¡: e ãã¹ãã£ã³ã°ã«ãããLDAãç¨ããé ç®éé¡ä¼¼åº¦ã®ç®åº, æ å ±å¦çå¦ä¼è«æèª, Vol.55, No.1, pp.91 - 104, 2014.
*4:Deerwester, S., Dumais, S. T., Furnas, G. W., Landauer, T. K. and Harshman, R.: Indexing by Latent Semantic Analysis, Vol. 41, No. 6, pp. 391â407 (1990).
*5:Hofmann, T.: Probabilistic latent semantic indexing, Proceedings of the 22nd annual international ACM SIGIR conference on Research and development in information retrieval, SIGIR â99, New York, NY, USA, pp. 50â57 (1999).
*6:Blei, D. M., Ng, A. Y. and Jordan, M. I.: Latent dirichlet allocation, J. Mach. Learn. Res., Vol. 3, pp. 993â1022 (2003).
*7:Teh, Y. W., Jordan, M. I., Beal, M. J. and Blei, D. M.: Hierarchical Dirichlet Processes, Journal of the American Statistical Association, Vol. 101, pp. 1566â1581(2006).
*8:Blei D. M. and Lafferty J. D.: Dynamic topic models, Proceedings of the 23rd international conference on Machine learning, pp. 113-120 (2006).
*9:Yan, X., Guo, J., Lan, Y. and Cheng, X.: A biterm topic model for short texts, Proceedings of the 22nd international conference on World Wide Web, pp. 1445â1456 (2013).
*10:ä¸å·è£å¿ï¼æ¹¯æ¬ç´å½°ï¼æ£®ãè¾°åï¼åºç¾é »åº¦ã¨é£æ¥é »åº¦ã«åºã¥ãå°éç¨èªæ½åºï¼èªç¶è¨èªå¦çï¼Vol. 10, No. 1, pp. 27â45 (2003).