ç§ããããèªãã çç±ã説æããããšã¯äžå¯èœã§ãã ã¡ããã©æéããã£ãã®ã§ãåžå Žãã©ã®ããã«æ©èœãããã«èå³ããããŸããã Gartner ã«ããã°ããã㯠2018 幎以æ¥ãã§ã«æ¬æ Œçãªåžå Žãšãªã£ãŠããŸãã 2014 幎ãã 2016 幎ãŸã§ã¯ã¢ããã³ã¹ã ã¢ããªãã£ã¯ã¹ (BI ã«ã«ãŒã) ãšåŒã°ããŠããŸãããã2017 幎ã«ã¯ããŒã¿ ãµã€ãšã³ã¹ (ããããã·ã¢èªã«ã©ãèš³ããããããããããŸãã) ãšåŒã°ããŠããŸããã åºå ŽåšèŸºã®åºåºè
ã®åãã«èå³ãããæ¹ã¯ã
ããã¯äœç³»çãªåæãè¡šã§ã¯ãããŸããã å°çç©çåŠè ã®èŠ³ç¹ããã®å人çãªèŠè§£ã ããããç§ã¯åžžã« Gartner MQ ãèªãããšã«èå³ãæã£ãŠããŸããMQ ã¯ããã€ãã®ç¹ãå®ç§ã«å®åŒåããŠããŸãã ããã§ãæè¡çãåžå ŽçããããŠå²åŠçãªèŠ³ç¹ããç§ã泚ç®ããç¹ã以äžã«ç€ºããŸãã
ããã¯ãML ã®ãããã¯ã«æ·±ãé¢ãã£ãŠãã人åãã§ã¯ãªããåžå Žã§äžè¬çã«äœãèµ·ãã£ãŠãããã«èå³ããã人åãã§ãã
DSML åžå Žèªäœã¯ãBI ãšã¯ã©ãŠã AI éçºè ãµãŒãã¹ã®éã«è«ççã«å ¥ãåã«ãªã£ãŠããŸãã
ãæ°ã«å
¥ãã®åŒçšãšçšèªãæåã«æããŸãã
- ããªãŒããŒã¯æè¯ã®éžæã§ã¯ãªããããããªãã â ããŒã±ãããªãŒããŒãå¿ ãããå¿ èŠãªããã§ã¯ãããŸããã éåžžã«ç·æ¥ïŒ æ©èœçãªé¡§å®¢ãããªãããã顧客ã¯åžžã«ãé©åãªããœãªã¥ãŒã·ã§ã³ã§ã¯ãªããæé©ãªããœãªã¥ãŒã·ã§ã³ãæ¢ããŠããŸãã
- ãã¢ãã«ã®éçšåã - MOP ãšç¥ãããŸãã ãããŠãã¿ããªãã°ã«ã¯èŠåŽããŠããŸãïŒ â (ã¯ãŒã«ãªãã°ã®ããŒããã¢ãã«ãæ©èœãããŸã)ã
- ãããŒãããœã³ã³ç°å¢ã ã¯ãã³ãŒããã³ã¡ã³ããããŒã¿ãçµæãçµã¿åããããéèŠãªæŠå¿µã§ãã ããã¯éåžžã«æ確ã§ææã§ãããUI ã³ãŒãã®éãå€§å¹ ã«åæžã§ããŸãã
- ããªãŒãã³ãœãŒã¹ã«æ ¹ãããã ããèšãããããšã§ããããªãŒãã³ãœãŒã¹ã«æ ¹ä»ããŠããŸãã
- ãã·ããºã³ããŒã¿ãµã€ãšã³ãã£ã¹ãã - èŠèŠç°å¢ãããããçš®é¡ã®è£å©çãªãã®ãå¿ èŠãšãããå°é家ã§ã¯ãªãããšãŠãç°¡åãªç·ããšãŠãäžæãªäººã 圌ãã¯ã³ãŒããæžããŸããã
- ãæ°äž»äž»çŸ©ã â å€ãã®å Žåããããå¹ åºã人ã ãå©çšã§ããããã«ããããšããæå³ã§äœ¿çšãããŸãã 以å䜿çšããŠããå±éºãªãããŒã¿ãèªç±ã«ãããã®ä»£ããã«ããããŒã¿ãæ°äž»åããããšèšããã§ãããã ãæ°äž»åãã¯åžžã«ãã³ã°ããŒã«ã§ããããã¹ãŠã®ãã³ããŒããããè¿œããããŸãã ç¥èã®éäžåã¯å€±ãããŸãããã¢ã¯ã»ã¹ããããã¯åäžããŸãã
- ãæ¢çŽ¢çããŒã¿åæ - EDAã â ãããã®å©çšå¯èœãªæ段ã®æ€èšã ããã€ãã®çµ±èšã ã¡ãã£ãšããèŠèŠåã å€ããå°ãªãã誰ãããã£ãŠããäºã ããã«ååããããšã¯ç¥ããŸããã§ãã
- ãåçŸæ§ã - äžåºŠå®æœããå®éšãç¹°ãè¿ãããšãã§ããããã«ããã¹ãŠã®ç°å¢ãã©ã¡ãŒã¿ãå ¥åããã³åºåãæ倧éã«ä¿åããã å®éšçãªãã¹ãç°å¢ãè¡šãæãéèŠãªçšèªã§ãã
ã ããïŒ
ã¢ã¬ãã¯ã¹
ãŸãã§ããã¡ãã®ãããªã¯ãŒã«ãªã€ã³ã¿ãŒãã§ãŒã¹ã ãã¡ãããã¹ã±ãŒã©ããªãã£ã¯å°ãé£ããã§ãã ãããã£ãŠãCitizen ã³ãã¥ããã£ã®ãšã³ãžãã¢ã¯ãã¡ãã£ãšããéã³ãããã®ãšåãã§ãã åæã¯ãã¹ãŠ XNUMX ã€ã®ããã«ã§è¡ããŸãã ã¹ãã¯ãã«çžé¢ããŒã¿è§£æã®è€éããæãåºããŸãã
ã¢ãã³ã³ã
Python ãš R ã®å°é家ã«é¢ããã³ãã¥ããã£ã ãªãŒãã³ãœãŒã¹ã¯ããã«å¿ããŠèŠæš¡ã倧ããã ç§ã®ååããã€ãããã䜿ã£ãŠããããšãããããŸããã ã§ããç¥ããŸããã§ããã
ããŒã¿ããªãã¯
ãã㯠2013 ã€ã®ãªãŒãã³ãœãŒã¹ ãããžã§ã¯ãã§æ§æãããŠããŸããSpark éçºè 㯠XNUMX 幎以æ¥ãè«å€§ãªè³éãéããŠããŸããããŠã£ããåŒçšããå¿ èŠããããŸãã
ã2013 幎 13.9 æãDatabricks 㯠Andreessen Horowitz ãã 33 äžãã«ã調éãããšçºè¡šããŸããã å瀟ã¯2014幎ã«60äžãã«ã2016幎ã«140äžãã«ã2017幎ã«250å2019äžãã«ã400幎ïŒ2019æïŒã«XNUMXåXNUMXäžãã«ãXNUMX幎ïŒXNUMXæïŒã«XNUMXåãã«ãããã«èª¿éããŸãããã
ã¹ããŒã¯ãã«ããããå人ãããŸãã ããããŸãããããããªããïŒ
ãããŠãããžã§ã¯ãã¯æ¬¡ã®ãšããã§ãã
- ãã«ã¿æ¹ - Spark äžã® ACID ãæè¿ãªãªãŒã¹ãããŸãã (Elasticsearch ã§ç§ãã¡ã倢èŠãŠãããã®ã§ã) - ãããããŒã¿ããŒã¹ã«å€ããŸã: å³æ Œãªã¹ããŒããACIDãç£æ»ãããŒãžã§ã³...
- ML ãã㌠â ã¢ãã«ã®è¿œè·¡ãããã±ãŒãžåã管çããã³ä¿ç®¡ã
- ã³ã¢ã© - Spark äžã® Pandas DataFrame API - Pandas - ããŒãã«ãããŒã¿å šè¬ãæäœããããã® Python APIã
Spark ãç¥ããªã人ããŸãã¯å¿ããŠããŸã£ã人ã®ããã«ã以äžãåç
§ããŠãã ããã
ã€ãŸããDatabricks 㯠Spark ãåŒãåºããŸãã Spark ãã¯ã©ãŠãã§éåžžã©ãã䜿çšããã人ã¯ãæå³ãããšãããããããããšãªã DataBricks ã䜿çšããŸã ð ããã§ã®äž»ãªå·®å¥åèŠå 㯠Spark ã§ãã
Spark Streaming ã¯æ¬ç©ã®åœã®ãªã¢ã«ã¿ã€ã ããã€ã¯ããããã§ã¯ãªãããšãåŠã³ãŸããã æ¬ç©ã®ãªã¢ã«ã¿ã€ã ãå¿
èŠãªå Žåã¯ãApache STORM ã䜿çšããŸãã ãŸãã誰ãã Spark ã MapReduce ãããåªããŠãããšèšããæžããŠããŸãã ãããã¹ããŒã¬ã³ã§ãã
ãã¿ã€ã¯
ãšã³ãããŒãšã³ãã®ã¯ãŒã«ãªãã®ã åºåããããããããŸãã Alteryx ãšã®éããããããŸããã
DataRobot
ããŒã¿æºåãæ åœãã Paxata ã¯ã2019 幎 20 æã« Data Robots ã«è²·åãããå¥äŒç€Ÿã§ãã 7äžãã«ã調éã売åŽããŸããã å šéšXNUMX幎ã§ã
Excel ã§ã¯ãªã Paxata ã§ã®ããŒã¿æºå - ãããåç
§ããŠãã ãã:
XNUMX ã€ã®ããŒã¿ã»ããéã®çµåã«ã€ããŠã¯ãèªåæ€çŽ¢ãšææ¡ãè¡ãããŸãã çŽ æŽãããããšã§ã - ããŒã¿ãç解ããã«ã¯ãããã¹ãæ
å ±ãããã«éèŠãããããšã«ãªããŸã (
Data Catalog ã¯ã圹ã«ç«ããªããã©ã€ããããŒã¿ã»ããã®åªããã«ã¿ãã°ã§ãã
Paxata ã§ãã£ã¬ã¯ããªãã©ã®ããã«åœ¢æãããã®ããèå³æ·±ãã§ã (
ãã¢ããªã¹ãäŒç€Ÿã«ãããšã
åµå ããœãããŠã§ã¢ã¯æè¡ã®é²æ©ã«ãã£ãŠå¯èœã«ãªããŸãããäºæž¬åæ ,æ©æ¢°åŠç¿ ãšNoSQL ããŒã¿ ãã£ãã·ã¥ææ³ãã15] ãœãããŠã§ã¢ã¯ã»ãã³ãã£ã㯠ããŒã¿ããŒãã«ã®åã®æå³ãç解ããã¢ã«ãŽãªãºã ãšãããŒã¿ã»ããå ã®æœåšçãªéè€ãèŠã€ãããã¿ãŒã³èªèã¢ã«ãŽãªãºã ã§ããã15] ã7] ãŸããã€ã³ããã¯ã¹äœæãããã¹ã ãã¿ãŒã³èªèããã®ä»ã®ãœãŒã·ã£ã« ã¡ãã£ã¢ãæ€çŽ¢ãœãããŠã§ã¢ã§äŒçµ±çã«äœ¿çšãããŠãããã¯ãããžãŒã䜿çšãããŠããŸããã
ããŒã¿ããããã®äž»å補åã¯ã
ãã¡ãããããŒã¿ ãµã€ãšã³ãã£ã¹ãã®å€§èŠæš¡ãªããŒã ã«ã¯ãã¢ãã«ãæäœããããã®ãŸãã«ãã®ãããªç°å¢ãå¿ èŠã§ããããšã¯æããã§ããããã§ãªããšã倧éã®ã¢ãã«ãäœæãããäœããããã€ãããªããªããŸãã ãããŠãç³æ²¹ãšã¬ã¹ã®äžæµã®çŸå®ã«ãããŠãæåããã¢ãã«ã XNUMX ã€äœæã§ããã°ãããã¯å€§ããªé²æ©ãšãªãã§ãããã
ãã®ããã»ã¹èªäœã¯ãããšãã°ãå°è³ªåŠãå°çç©çåŠã«ãããèšèšã·ã¹ãã ã®ç 究ãéåžžã«æãåºãããŸãã
ããã
ãªãŒãã³ãã©ãããã©ãŒã ãšã³ã©ãã¬ãŒã·ã§ã³ãéèŠããŸãã ããžãã¹ãŠãŒã¶ãŒã¯ç¡æã§ãå©çšããã ããŸãã 圌ãã® Data Lab 㯠SharePoint ã«éåžžã«äŒŒãŠããŸãã (ãããŠãã®ååã¯IBMã匷ãåãããŸã)ã ãã¹ãŠã®å®éšã¯å ã®ããŒã¿ã»ããã«ãªã³ã¯ããŠããŸãã ããã¯ããããããšã§ãã:) ç§ãã¡ã®å®è·µãšåãããã«ãããã€ãã®ããŒã¿ãã¢ãã«ã«ãã©ãã°ããããã®åŸã¯ãªãŒã³ã¢ãããããã¢ãã«å ã§æŽçãããŸãããããããã¹ãŠã¯ãã§ã«ã¢ãã«å ã«ååšããŠããããœãŒã¹ããŒã¿ã§ã¯çµãããèŠã€ãããŸããã ã
Domino ã«ã¯åªããã€ã³ãã©ã¹ãã©ã¯ãã£ä»®æ³åæ©èœããããŸãã ç§ã¯ããã«å¿ èŠãªæ°ã®ã³ã¢ããã·ã³ã«çµã¿ç«ãŠãæ°ãã«è¡ããŸããã ãããã©ã®ããã«è¡ãããã®ãã¯ããã«ã¯æããã§ã¯ãããŸããã Docker ã¯ã©ãã«ã§ãååšããŸãã èªç±åºŠãã£ã·ãïŒ ææ°ããŒãžã§ã³ã®ã¯ãŒã¯ã¹ããŒã¹ã§ããã°æ¥ç¶å¯èœã§ãã 䞊è¡ããŠå®éšãéå§ã æåãããã®ã®è¿œè·¡ãšéžæã
DataRobot ãšåã - çµæã¯ã¢ããªã±ãŒã·ã§ã³ã®åœ¢åŒã§ããžãã¹ ãŠãŒã¶ãŒåãã«å ¬éãããŸãã ç¹ã«æèœã®ãããé¢ä¿è ãåãã ãŸããã¢ãã«ã®å®éã®äœ¿çšç¶æ³ãç£èŠãããŸãã ãã¹ãŠã¯ãã°ã®ããã«ïŒ
è€éãªã¢ãã«ãã©ã®ããã«ããŠæ¬çªç°å¢ã«å°å ¥ãããã®ããå®å šã«ã¯ç解ã§ããŸããã ããŒã¿ããã£ãŒãããŠçµæãååŸããããã«ããã皮㮠API ãæäŸãããŠããŸãã
H2O
Driveless AI ã¯ãæåž«ãã ML çšã®éåžžã«ã³ã³ãã¯ãã§çŽæçãªã·ã¹ãã ã§ãã ãã¹ãŠã XNUMX ã€ã®ããã¯ã¹ã«åãŸããŸãã ããã¯ãšã³ãã«ã€ããŠã¯ãããã«ã¯å®å šã«æããã§ã¯ãããŸããã
ã¢ãã«ã¯ãREST ãµãŒããŒãŸã㯠Java ã¢ããªã«èªåçã«ããã±ãŒãžåãããŸãã ããã¯çŽ æŽãããã¢ã€ãã¢ã§ãã 解éå¯èœæ§ãšèª¬æå¯èœæ§ã®ããã«å€ãã®ããšãè¡ãããŠããŸããã ã¢ãã«ã®çµæã®è§£éãšèª¬æ (æ¬è³ªçã«èª¬æã§ããªããã®ã¯äœããããã§ãªããã°äººéãåãããšãèšç®ã§ããã®ã)ã
åããŠãéæ§é åããŒã¿ãš
å®å
šã«æ確ã§ã¯ãªã倧èŠæš¡ãªãªãŒãã³ãœãŒã¹ H2O ãã¬ãŒã ã¯ãŒã¯ããããŸã (ã¢ã«ãŽãªãºã /ã©ã€ãã©ãªã®ã»ãã?)ã Jupiter ã®ãããªããã°ã©ãã³ã°ãå¿
èŠãšããªãç¬èªã®ããžã¥ã¢ã« ã©ããããã (
åãå Žæã§: ããŒããŠã§ã¢ãšã¯ã©ãŠããšã®çµ±ååéã«ãããé«æ§èœãæé©åãæ¥çæšæºã
ãããŠããã®åŒ±ç¹ã¯è«ççã§ã - Driverles AI ã¯ããªãŒãã³ãœãŒã¹ãšæ¯èŒããŠåŒ±ããç¯å²ãçãã§ãã Paxataã«æ¯ã¹ãŠããŒã¿æºåãããµãïŒ ãããŠãã¹ããªãŒã ãã°ã©ããå°çæ å ±ãªã©ã®ç£æ¥ããŒã¿ãç¡èŠããŸãã ãŸãããã¹ãŠãè¯ãããšã°ããã§ã¯ãããŸããã
éšå£«
ã¡ã€ã³ããŒãžã«ãã 6 ã€ã®éåžžã«å ·äœçã§èå³æ·±ãããžãã¹ ã±ãŒã¹ãæ°ã«å ¥ããŸããã 匷åãªãªãŒãã³ãœãŒã¹ã
Gartner ã¯åœŒãããªãŒããŒããããžã§ããªãŒã«éæ ŒãããŸããã ãªãŒããŒãåžžã«æè¯ã®éžæã§ãããšã¯éããªããããåçãäžååã§ããããšã¯ãŠãŒã¶ãŒã«ãšã£ãŠè¯ãå åã§ãã
ããŒã¯ãŒãã¯ãH2O ãšåæ§ã«ãæ¡åŒµã§ããããã¯ã貧ããåžæ°ã®ããŒã¿ ãµã€ãšã³ãã£ã¹ããæ¯æŽããããšãæå³ããŸãã ã¬ãã¥ãŒã§ããã©ãŒãã³ã¹ã«ã€ããŠæ¹å€ãããã®ã¯ãããåããŠã§ãã é¢çœãïŒ ã€ãŸããéåžžã«å€ãã®ã³ã³ãã¥ãŒãã£ã³ã°èœåããããããããã©ãŒãã³ã¹ãã·ã¹ãã äžã®åé¡ã«ãªãå¯èœæ§ã¯ãŸã£ãããããŸããã Gartner ã¯ãã®ãæ¡åŒµããšããèšèã«ã€ããŠæ¬¡ã®ããã«è¿°ã¹ãŠããŸãã
ãããŠãKNIMEã¯ã¬ãã¥ãŒã§æåã®éã¢ã¡ãªã«äººã®ããã§ãïŒ (ãããŠãç§ãã¡ã®ãã¶ã€ããŒã¯ã©ã³ãã£ã³ã° ããŒãžããšãŠãæ°ã«å
¥ã£ãŠããŸãããå¥åŠãªäººãã¡ã§ãã
MathWorks
MatLab ã¯èª°ããç¥ã£ãŠããå€ãåèªåå¿ã§ãã ç掻ã®ããããåéãç¶æ³ã«å¯Ÿå¿ããããŒã«ããã¯ã¹ã äœããšãŠãéãã å®éã人çã®ãããããã®ã«ã¯ãéåžžã«å€ãã®æ°åŠãå¿ èŠã§ãã
ã·ã¹ãã èšèšçšã® Simulink ã¢ããªã³è£œåã ããžã¿ã«ãã€ã³ã®ããŒã«ããã¯ã¹ã調ã¹ãŸãã - ããã«ã€ããŠã¯äœãç解ããŠããŸãããã
RapidMiner
ç§ã¯ãããŸã§ãåªãããªãŒãã³ãœãŒã¹ã®æè㧠(Matlab ãšãšãã«) ããããã®ããšã«åºäŒã£ããèãããããŠããŸããã ãã€ãã®ããã«TurboPrepã«ã€ããŠå°ã調ã¹ãŠã¿ãŸããã ããŒãã£ããŒã¿ããã¯ãªãŒã³ããŒã¿ãååŸããæ¹æ³ã«èå³ããããŸãã
ããã§ãã2018 幎ã®ããŒã±ãã£ã³ã°è³æãšãæ©èœãã¢ã§è±èªã話ã人ã ãã²ã©ãããšããã人ã ãè¯ãããšãããããŸãã
2001 幎以éãã«ãã ã³ãåºèº«ã§ã匷ããã€ãã®èæ¯ãæã€äººã ïŒ
ãã®ãµã€ããèŠãã ãã§ã¯ããªãŒãã³ãœãŒã¹ã§äœãå©çšã§ããã®ããŸã ç解ã§ããŸãããããã«è©³ãã調ã¹ãå¿
èŠããããŸãã å°å
¥ãš AutoML ã®æŠå¿µã«é¢ããåªãããããªã
RapidMiner Server ããã¯ãšã³ãã«ã€ããŠãç¹å¥ãªããšã¯äœããããŸããã ããããã³ã³ãã¯ãã§ãç®±ããåºããŠããã«ãã¬ãã¢ã ã§ããŸãæ©èœããã§ãããã Docker ã«ããã±ãŒãžåãããŠããŸãã RapidMiner ãµãŒããŒäžã®ã¿ã®å ±æç°å¢ã ãããŠãRadoopãHadoop ããã®ããŒã¿ãStudio ã¯ãŒã¯ãããŒã® Spark ããé»ãæ°ããŸãã
äºæ³éããè¥ã人æ°ã®ãã³ããŒãçžæ£ã®å£²ãæããããããäžã«ç§»åãããŸããã ããããGartner ã¯ãšã³ã¿ãŒãã©ã€ãºåéã§ã®å°æ¥ã®æåãäºæž¬ããŠããŸãã ããã§ãéãéããããšãã§ããŸãã ãã€ã人ã¯ãããè¡ãæ¹æ³ãç¥ã£ãŠããŸãããªããšã :) SAP ã«ã€ããŠã¯èšåããªãã§ãã ããã
圌ãã¯åœæ°ã®ããã«ããããã®ããšãããŠãããŠããŸãïŒ ãããããã®ããŒãžãèŠããšãGartner ã¯è²©å£²é©æ°ã«èŠæŠããŠãããã«ããŒç¯å²ã®åºãã§ã¯ãªãåçæ§ãæ±ããŠæŠã£ãŠãããšè¿°ã¹ãŠããããšãããããŸãã
æ®ã£ã SAS О ãã£ãã³ ç§ã«ãšã£ãŠå
žåç㪠BI ãã³ããŒã§ã...ãããŠäž¡æ¹ãšãæäžäœã«ãããããã¯éåžžã®ããŒã¿ãµã€ãšã³ã¹ãè«ççã«æé·ããŠãããšããç§ã®èªä¿¡ãè£ä»ããŠããŸã
ã¯ã©ãŠãã Hadoop ã€ã³ãã©ã¹ãã©ã¯ãã£ããã§ã¯ãªããBI ããã ã€ãŸããIT ããã§ã¯ãªããããžãã¹ããã§ãã ããšãã°ã¬ã¹ããã ãããã®ããã«ïŒ
SAS
èšãããšã¯ããŸããããŸããã æãããªããšã ãã
TIBCO
ãã®æŠç¥ã¯ã28 ããŒãžã«ããã Wiki ããŒãžã®è²·ãç©ãªã¹ãã«èšèŒãããŠããŸãã ã¯ãã話ã¯é·ããªããŸããã2007 !!! ãã£ãŒã«ãºã ç§ã¯ãã¯ãéæ¥æ代㫠BI Spotfire (2014) ãè³Œå ¥ããŸããã ãŸããJaspersoft (2008)ããã®åŸ 2017 瀟ãã®äºæž¬åæãã³ã㌠Insightful (S-plus) (2017)ãStatistica (2013)ãAlpine Data (2018)ãã€ãã³ãåŠçããã³ã¹ããªãŒãã³ã° Streambase System (2019)ãMDM Orchestra ãããã¬ããŒããæäŸãããŠããŸãã Networks (XNUMX) ãš Snappy Data (XNUMX) ã®ã€ã³ã¡ã¢ãª ãã©ãããã©ãŒã ã
ããã«ã¡ã¯ããã©ã³ããŒïŒ
åºæïŒ habr.com