¼ÆËã»úÓ¦Óü¼ÊõÂÛÎÄ

Äúµ±Ç°µÄλÖãºÑ§ÊõÌà > ¼ÆËã»úÂÛÎÄ > ¼ÆËã»úÓ¦Óü¼ÊõÂÛÎÄ >

ÄÌÅ£¼²²¡Ô¤²âÁìÓòÖлúÆ÷ѧϰµÄÓ¦ÓÃ

À´Ô´£º¶¯Îïҽѧ½øÕ¹ ×÷Õߣº·ëåû,¸ßÖ¾Ìì
·¢²¼ÓÚ£º2021-07-16 ¹²7647×Ö
¡¡¡¡±¾ÆªÂÛÎÄ¿ìËÙµ¼º½£º

Õ¹¿ª¸ü¶à

»úÆ÷ѧϰÂÛÎÄ·¶ÎĵÚËÄƪ£ºÄÌÅ£¼²²¡Ô¤²âÁìÓòÖлúÆ÷ѧϰµÄÓ¦ÓÃ

¡¡¡¡ÕªÒª£º»úÆ÷ѧϰÊÇÖ¸¼ÆËã»úͨ¹ý´óÁ¿Êý¾ÝѵÁ·¼°·ÖÎöÀ´Ä£ÄâÈËÀàµÄѧϰÐÐΪ£¬´Ó¶ø»ñµÃеÄ֪ʶºÍ¼¼ÄÜ£¬ÊÇÈ˹¤ÖÇÄܵĺËÐÄ¡£½üÄêÀ´£¬»úÆ÷ѧϰÔÚÄÌÅ£¼²²¡Ô¤²âÁìÓòµÄÓ¦ÓÃÒѳÉΪ¹ú¼ÊÉϵÄÑо¿Èȵ㡣ÂÛÎĽéÉÜÁËÀûÓÃÄÌÅ£»úÌåÉúÀíÖ¸±êºÍÉú²úÊý¾Ý½¨Á¢ÄÌÅ£¼²²¡Ô¤²âÄ£Ð͵ķ½·¨£¬Öصã²ûÊöÁËÈçºÎÀûÓþö²ßÊ÷ºÍÉñ¾­ÍøÂçɸѡÄÌÅ£¼²²¡·çÏÕÒò×Ó¡¢Ô¤²â¼²²¡ºÍ¼²²¡·ÖÀࡣͬʱ£¬×ÛÊöÁË»úÆ÷ѧϰԤ²â´úлÐÔ¼²²¡¡¢õËÐС¢Èé·¿Ñס¢ÈÈÓ¦¼¤ºÍ´«È¾ÐÔ¼²²¡µÄ½øÕ¹¡£

¡¡¡¡¹Ø¼ü´Ê£º»úÆ÷ѧϰ£»ÄÌÅ££»ÁÙ´²¼²²¡£»¾ö²ßÊ÷£»Éñ¾­ÍøÂ磻

¡¡¡¡×÷Õß¼ò½é£º·ëåû£¨1978-£©£¬Å®£¬ÉÂÎ÷µ¤·ïÈË£¬½²Ê¦£¬²©Ê¿£¬Ö÷Òª´ÓÊÂÎïÁªÍøÅ©ÒµÓ¦ÓúͻúÆ÷ѧϰÑо¿¡££»

¡¡¡¡Abstract:The concept of machine learning refers to a computer that simulates human learning behavior through a large amount of data training and analysis to obtain new knowledge and skills.Machine learning is the core of artificial intelligence.In recent years, the research of machine learning in the field of dairy cow disease prediction has become a hot topic in the world.This paper introduced the method of establishing dairy cow disease prediction model by using physiological indicators and production data of dairy cow.The approach to use decision trees and neural networks to select disease risk factors, predict diseases and classify diseases was stressed.At the same time, the progress of machine learning in predicting metabolic diseases, lameness, mastitis, heat stress and infectious diseases was reviewed.

¡¡¡¡Keyword:Machine learning; dairy cow; clinical disease; decision tree; neural network;

»úÆ÷ѧϰ

¡¡¡¡»úÆ÷ѧϰ£¨Machine learning, ML£©ÊÇÑо¿ÈçºÎʹ»úÆ÷ͨ¹ýʶ±ðºÍÀûÓÃÏÖÓÐ֪ʶ»ñÈ¡ÐÂ֪ʶµÄÒ»ÃŶàÁìÓò½»²æѧ¿Æ¡£ML¿Éͨ¹ýѧϰÒÑÓÐÊý¾Ý£¬½¨Á¢Ò»ÖÖÄ£ÐÍ»òѧϰÆ÷£¬¶Ôδ֪µÄÊý¾Ý½øÐзÖÎöºÍÔ¤²â¡£µ±½ñµÄÅ©ÒµÉú²ú»á²úÉú´óÁ¿µÄÊý¾Ý£¬ÀûÓÃML¼¼Êõ·ÖÎöÕâЩÊý¾Ý²¢½¨Ä£½«ÊÇÅ©Òµ´óÊÆËùÇ÷[1].ÔÚÄÌÅ£Òµ£¬Ëæ×ÅÄÌÅ£ÑøÖ³µÄ¾«Ï¸»¯ºÍÈ˹¤³É±¾µÄÔö¼Ó£¬½ö¿¿Å£³¡µÄ¹ÜÀíÈËÔ±ÎÞ·¨Íê³É¸ßÖÊÁ¿µÄȺÌ廯ºÍ¸öÐÔ»¯¹ÜÀí¹¤×÷£¬Ê¹MLÔÚÄÌÅ£¾«×¼ÓýÖÖ¡¢ÈºÌå¹ÜÀíºÍ¼²²¡¼à²âµÈÁìÓòÓÐ׏㷺µÄÑо¿ºÍÓ¦ÓÃÇ°¾°[2,3].ÔÚÕâЩÁìÓòÖУ¬ÀûÓÃMLË㷨ѧϰºÍѵÁ·ÄÌÅ£¼²²¡µÄÌØÓзçÏÕÒòËØÒÔÔ¤²âºÍ¼à²âÄÌÅ£¼²²¡ÒѳÉΪ¾«×¼ÈéÒµµÄÑо¿·½Ïò[4,5,6].Ó¦ÓÃMLÔ¤²âÄÌÅ£´úлÐÔ¼²²¡¡¢Èé·¿Ñס¢´«È¾²¡¡¢ÈÈÓ¦¼¤ºÍõËÐеȼ²²¡ÒѳÉΪ¹ú¼ÊÉϵÄÑо¿Èȵã[4,5,6,7,8,9,10],µ«ÎÒ¹úÔÚ´ËÁìÓòµÄÑо¿Æð²½½ÏÍí¡£±¾ÎľÍMLÔÚÄÌÅ£ÁÙ´²¼²²¡Ô¤²â·½ÃæµÄÑо¿×öÒ»×ÛÊö£¬ÎªÎÒ¹ú¿ªÕ¹Õâ·½ÃæµÄÑо¿Ìṩ²Î¿¼¡£

¡¡¡¡1 Ñо¿·½·¨ºÍÔ­Àí

¡¡¡¡1.1 ML¼ò½é

¡¡¡¡MLÀûÓÃÊýѧ·½·¨ºÍ¼ÆËã»ú¼¼Êõ¶ÔÀúÊ·Êý¾Ý½øÐзÖÎöµÃµ½¹æÂɲ¢¹¹½¨Ä£ÐÍ£¬¶Ôδ֪Êý¾Ý½øÐÐÔ¤²âºÍ·ÖÀà¡£½¨Ä£Ê±£¬ÏȽ«Ô¤´¦ÀíºÃµÄÄÌÅ£Êý¾ÝÊäÈëµ½ML·ÖÀàÆ÷½øÐÐѵÁ·£¬È»ºóµÃµ½¶Ôδ֪Êý¾Ý½øÐÐÔ¤²âµÄÄ£ÐÍ£¬¶ÔÔ¤²â½á¹û¾ßÓнϴó¹±Ï×µÄÌØÕ÷¿É×÷ΪÄÌÅ£»¼²¡µÄDZÔÚÌØÕ÷[11].Õâ¸ö¹ý³Ì°üÀ¨Êý¾ÝÊÕ¼¯¡¢Êý¾ÝÔ¤´¦Àí¡¢Ä£ÐÍѵÁ·¡¢Ä£ÐÍÑ¡ÔñºÍ½á¹ûÔ¤²â£¬¾ßÌ彨ģÁ÷³ÌÈçͼ1Ëùʾ¡£³£ÓõÄML±à³ÌÓïÑÔÓÐPython, R,MATLABºÍOctaveµÈ£¬¾ùÓд¦ÀíÊý¾ÝµÄͳ¼ÆÈí¼þ°ü£¬Ê¹·ÖÎöÊý¾Ý±äµÃÈÝÒס£

¡¡¡¡1.2 Êý¾Ý´¦Àí

¡¡¡¡ÓÃÓÚÄÌÅ£¼²²¡¼à²âµÄÊý¾Ýͨ³£°üÀ¨Å©³¡´æ´¢µÄÊý¾Ý¡¢ÉúÀíÌØÕ÷Êý¾Ý¡¢Î¹Ñø¹ý³ÌÖеÄÊý¾Ý¡¢ËÇÑø¹ÜÀíÈí¼þÊý¾ÝºÍ´«¸ÐÆ÷²É¼¯µÄÊý¾ÝµÈ¡£ÔÚʵ¼ÊÓ¦ÓÃÖУ¬ÒòΪÒÔÉÏÊý¾Ý´æÔÚ²»ÍêÕû¡¢Öظ´¡¢È˹¤Â¼Èë´íÎóºÍÊýֵȱʧµÈÇé¿ö£¬ÎªÁË»ñµÃ¸üºÃµÄÔ¤²â½á¹û£¬±ØÐë¶ÔÊý¾Ý×öÔ¤´¦Àí¡£Êý¾ÝÔ¤´¦Àíͨ³£ÏȽøÐÐÊý¾ÝÇåÏ´¡¢Êý¾Ý±ä»»¡¢Êý¾Ý¼¯³ÉºÍÊý¾Ý¼ò»¯µÈ£¬ÔÙ½«Êý¾Ý¼¯»®·ÖΪѵÁ·¼¯ºÍ²âÊÔ¼¯Á½²¿·Ö¡£ÆäÖÐѵÁ·¼¯ÓÃÓÚÄ£Ð͵ÄѵÁ·ºÍ¹¹½¨£¬²âÊÔ¼¯ÓÃÓÚÔ¤²â½á¹ûµÄÆÀ¹À£¬±ÈÀýÒ»°ãΪ70%ºÍ30%,ºÜ¶àʱºò»¹»á¶à»®·Ö³öÑéÖ¤¼¯£¨60%,20%ºÍ20%£©[11].

¡¡¡¡1.3 MLÔÚÄÌÅ£¼²²¡Ô¤²âÖеij£ÓÃËã·¨

¡¡¡¡ML°üÀ¨Óмලѧϰ¡¢·Ç¼à¶½Ñ§Ï°¡¢°ë¼à¶½Ñ§Ï°ºÍÇ¿»¯Ñ§Ï°£¬ML³£ÓõÄËã·¨·ÖÀàÈçͼ2Ëùʾ¡£¼à¶½Ñ§Ï°ÐèÒª±ê¼ÇÊý¾ÝºÍѵÁ·Êý¾Ý¼¯£¬¶ø·Ç¼à¶½Ñ§Ï°ÔÚδ±ê¼ÇÊý¾ÝµÄÇé¿ö϶ÀÁ¢ÆÀ¹ÀÊý¾Ý£»°ë¼à¶½Ñ§Ï°·½·¨Ê¹ÓõÄÊý¾Ý¼¯Ö»ÓÐһС²¿·Ö±ê¼ÇÊý¾Ý¡£Ó¦ÓÃMLÔ¤²âÄÌÅ£¼²²¡µÄÖ÷ÒªËã·¨ÓлùÓÚÊ÷Ä£ÐÍËã·¨¡¢È˹¤Éñ¾­ÍøÂçËã·¨£¨Artificial neural network, ANN£©£¨¼ò³ÆÉñ¾­ÍøÂ磩¡¢»Ø¹éËã·¨ºÍ¾ÛÀàËã·¨µÈ¡£Ê¹ÓÃ×îΪ¹ã·ºµÄÊÇ»ùÓÚÊ÷Ä£ÐÍËã·¨£¬±ÈÈç¾ö²ßÊ÷£¨Decision tree, DT£©ºÍËæ»úÉ­ÁÖ£¨Random forest, RF£©¡£ANNÔÚ¸÷ÀàÓ¦ÓÃÖоùÓÐʹÓã¬ÁíÍâ³£ÓõÄËã·¨»¹ÓÐÆÓËر´Ò¶Ë¹£¨naive bayes, NB£©¡¢K×î½üÁÚËã·¨£¨K-nearest neighbor, KNN£©ºÍÖ§³ÖÏòÁ¿»ú£¨support vector machine, SVM£©µÈ[12].

¡¡¡¡Í¼1 »úÆ÷ѧϰ½¨Ä£Á÷³Ì

¡¡¡¡Fig.1 The flowchart of ML modeling

¡¡¡¡Í¼2 »úÆ÷ѧϰ³£ÓÃËã·¨·ÖÀà

¡¡¡¡Fig.2 The classification of commonly used algorithms for ML

¡¡¡¡1.3.1 »ùÓÚÊ÷Ä£ÐÍËã·¨

¡¡¡¡Ô¤²âÄÌÅ£¼²²¡»ùÓÚÊ÷Ä£Ð͵ÄËã·¨Ö÷ÒªÓÐDT[13]¡¢RF[5]ºÍÌݶÈÌáÉý¾ö²ßÊ÷£¨gradient boosting decision tree, GBDT£©µÈ¡£DTÊÇÒ»¸öµ¹Á¢µÄÊ÷Ðνṹ£¬Ëã·¨ÀûÓÃѵÁ·¼¯´ÓÊ÷µÄ¸ù½áµã¿ªÊ¼£¬Í¨¹ý¶Ô¼²²¡¹Ø¼üÖ¸±ê½øÐÐ×Ô¶¥ÏòÏÂÒÀ´ÎµÄ¶¨ÐÔÅжÏÀ´Õï¶ÏºÍÔ¤²â¼²²¡¡£DTÔÚʹÓÃʱËٶȿ죬¾«¶È¸ß£¬Ä£Ðͼòµ¥Ò׶®£¬Òò´ËÔÚʵ¼ÊÔ¤²âÖÐʵÓÃÐÔÇ¿£¬Ð§¹ûºÃ[11].´ËÍ⣬RFÒò¿É¸ÄÉÆDTÒ×ÄâºÏµÄÌصãÒ²±»¹ã·ºÓ¦Óá£

¡¡¡¡1.3.2 È˹¤Éñ¾­ÍøÂç

¡¡¡¡ANNÊÇÒ»ÖÖÄ£ÄâÉúÎïÉñ¾­ÍøÂç½øÐÐÐÅÏ¢´¦ÀíµÄÊýѧģÐÍ£¬ÓÉÊäÈë²ã¡¢Òþº¬²ãºÍÊä³ö²ã×é³É¡£³£ÓõÄANNËã·¨Óз´Ïò´«²¥£¨back propagation, BP£©Éñ¾­ÍøÂç¡¢¾¶Ïò»ùº¯Êý£¨radial basis function, RBF£©Éñ¾­ÍøÂ硢ģºýÉñ¾­ÍøÂçºÍ×ÔÊÊÓ¦Éñ¾­ÍøÂçµÈ£¬ÆäÖÐBPÉñ¾­ÍøÂçºÍRBFÉñ¾­ÍøÂçÓÉÓÚÆäÁ¼ºÃµÄ·ÇÏßÐԱƽüÄÜÁ¦ÔÚ¼²²¡Ô¤²âÖб»¹ã·ºÓ¦ÓÃ[14,15].

¡¡¡¡1.3.3 ¾ÛÀà·ÖÎöËã·¨

¡¡¡¡¾ÛÀà·ÖÎöÊÇÖ¸Ñù±¾ÔÚûÓзÖÀà±ê×¼µÄÇé¿öÏ£¬¸ù¾ÝÑù±¾±¾ÉíµÄÊôÐÔ£¬ÓÃÊýѧͳ¼Æ·½·¨°´ÕÕijÖÖÏàËÆÐÔ»ò²îÒìÐÔÌØÕ÷£¬¶¨Á¿È·¶¨Ñù±¾Ö®¼äµÄÏàËƶȻò¾àÀ룬²¢°´ÕÕÕâÖÖÏàËƶȻò¾àÀëµÄÔ¶½ü¹Øϵ¶ÔÑù±¾½øÐоÛÀ࣬´Ó¶øɸѡ³ö¼²²¡µÄÌØÕ÷¡£³£ÓõľÛÀàËã·¨ÓÐK¾ùÖµ£¨K-means£©¾ÛÀࡢģºýC¾ùÖµ¾ÛÀàºÍ²ã´Î¾ÛÀà¡£¾ÛÀà·ÖÎöÄ£ÐÍÖ±¹Û¼òµ¥£¬ÊÊÓÃÓÚ´¦Àí¶àÖÖ²»Í¬ÀàÐ͵ÄÑù±¾Á¿Ð¡µÄÊý¾Ý¼¯ºÏ£¬ÔÚÄÌÅ£õËÐÐÕï¶Ï[16]¡¢´úлÐÔ¼²²¡[17]ºÍ´«È¾²¡µÄÔ¤²â[18]¾ùÓÐÓ¦Óá£

¡¡¡¡1.3.4 »Ø¹éËã·¨

¡¡¡¡»Ø¹éËã·¨ÊÇÒ»ÖÖ¶ÔÊýÖµÐÍÁ¬ÐøËæ»ú±äÁ¿½øÐÐÔ¤²âºÍ½¨Ä£µÄ¼à¶½Ñ§Ï°Ëã·¨¡£»Ø¹éËã·¨Òª½¨Á¢È·¶¨¼²²¡Ïà¹ØµÄΣÏÕÒòËغͼ²²¡Ö®¼äµÄÓ³Éä¹ØϵµÄº¯Êý£¬Ê¹µÃ²ÎÊýÖ®¼äµÄ¹ØϵÄâºÏÐÔ×îºÃ£¬µÃµ½Óë¼²²¡Ïà¹ØµÄΣÏÕÒòËØ£¬´Ó¶øɸѡ³öÔ¤²â¼²²¡µÄÊäÈë±äÁ¿£¬½¨Á¢×îÓŻع鷽³Ì½øÐн¨Ä£¡£³£ÓõĻعéËã·¨ÓÐÏßÐԻعéºÍÂß¼­»Ø¹é£¬ÏßÐԻعéÖ÷ÒªÓÃÀ´½â¾öÁ¬ÐøÖµÔ¤²âµÄÎÊÌ⣬Âß¼­»Ø¹éÓÃÀ´½â¾ö·ÖÀàµÄÎÊÌâ[11].

¡¡¡¡1.4 Ä£ÐÍÆÀ¼Û

¡¡¡¡Ñù±¾¼¯ÑµÁ·Íê³ÉºóÐèÒªÆÀ¹ÀÔ¤²âÄ£Ð͵ÄÐÔÄÜ£¬³£ÓõÄÆÀ¼ÛÖ¸±êÓÐ׼ȷÂÊ¡¢¾«¶ÈºÍÕÙ»ØÂʵȡ£

¡¡¡¡1.5 ÐÔÄÜÓÅ»¯

¡¡¡¡ÔÚÌØÕ÷ºÍÄ£ÐÍÈ·¶¨ºóͨ¹ýµ÷ÕûÄ£Ð͵IJÎÊýÀ´Ìá¸ß¼²²¡Ô¤²âÄ£Ð͵Ä׼ȷÂÊ¡£³£Óõĵ÷²Î·½·¨ÓÐÊÖ¹¤ËÑË÷¡¢Íø¸ñËÑË÷¡¢Ëæ»úËÑË÷ºÍ±´Ò¶Ë¹ËÑË÷µÈ¡£

¡¡¡¡2 MLÔÚÄÌÅ£ÁÙ´²¼²²¡ÖеÄÑо¿½øÕ¹

¡¡¡¡2.1 ´úлÐÔ¼²²¡

¡¡¡¡MLÔ¤²â´úл²¡Ê±£¬Ê×ÏȲɼ¯ÄÌÅ£ÉúÀíÖ¸±êºÍÉú²úÊý¾Ý£¬ÔÙ¸ù¾ÝÐèÇóɸѡ¿É·ÖÎöÉúÀíÖ¸±ê»òÉú²úÊý¾ÝºÍ¼²²¡Ö®¼ä¹ØϵµÄÄ£ÐÍ£¬×îºó½¨Ä£¡£±ÈÈ磬ΪÁËÑо¿¸ÉÄÌÆÚ¡¢Ì¥´Î¡¢²úÄÌÐÔ×´ºÍÌåÖصÈÊý¾ÝÄÜ·ñÔ¤²â²úºó´úл״̬£¬Xu WµÈ[5]±È½ÏÑо¿ÁËDT¡¢NB¡¢±´Ò¶Ë¹ÍøÂç¡¢SVM¡¢ANN¡¢KNN¡¢Bootstrap¾ÛºÏºÍËæ»úËÑË÷µÈ8ÖÖËã·¨µÄ½¨Ä£Ð§¹û£¬·¢ÏÖRFºÍSVMÔ¤²âЧ¹û½ÏºÃ¡£ÔÚ´úл²¡Ô¤²â·½Ã棬ͨ¹ýANN¶Ô»ùÒò×éºÍ´úлÐÅÏ¢½¨Ä££¬ÔÚ²úºó1¡¢3¡¢4¡¢5ÖÜÄܽÏΪ׼ȷԤ²âÑÇÁÙ´²Íª²¡[14].ɸѡΣÏÕÒòËØʱ£¬DTºÍRFËã·¨ÓÐÔËËã¿ìµÄÓÅÊÆ£¬Í¨¹ý·ÖÎöÄÌÅ£½¡¿µ¼Ç¼¾Í¿Éɸѡ³öµ¼ÖÂÌÔÌ­µÄÔ­ÒòÒÀ´ÎΪÈéÈÈ¡¢Öåθ±äλ¡¢ÁÙ´²ÈéÏÙÑס¢×Ó¹¬Ñ׺ÍË«Ì¥¡£²¢ÇÒ£¬µ±Á½¸ö¼²²¡µþ¼Óʱ£¬ÌÔÌ­µÄ·çÏÕ»á½øÒ»²½Ôö¼Ó[14].ÈôÒª·ÖÎöѪҺ´úл²úÎïºÍ¼²²¡Ö®¼äµÄ¹Øϵ£¬MLÐè½áºÏͳ¼Æѧ·½·¨½øÐоÛÀà¡¢»Ø¹é»ò½¨Á¢ÏßÐÔÄ£ÐÍ¡£Tremblay MµÈ[7]ͨ¹ýÖ÷³É·Ö·ÖÎöºÍK-means¾ÛÀà·ÖÎö£¬·¢ÏÖ·Çõ¥»¯Ö¬·¾ËáˮƽÓë´úлÊÊÓ¦²»Á¼×ÛºÏÕ÷ÏÔÖøÕýÏà¹Ø¡£Van Hoeij RµÈ[17]ÀûÓùãÒåÏßÐÔÄ£ÐÍ·ÖÎö·¢ÏÖ´úл״¿ö²îµÄÄÌÅ£¸ÉÎïÖÊÉãÈëÁ¿µÍ£¬Ò×·¢ÉúÄÜÁ¿¸ºÆ½ºâ¡£

¡¡¡¡2.2 õËÐÐ

¡¡¡¡ML·ÖÎöõËÐÐʱ£¬Ö÷ÒªÊǶÔÄÌÅ£Ô˶¯Ê±²»Í¬²¿Î»µÄÔ˶¯Í¼Ïñ½øÐв½Ì¬·ÖÎö£¬É¸Ñ¡³öÌØÕ÷²ÎÊýºó½¨Ä££¬×îºó½øÐзÖÀàÑéÖ¤¡£²ÉÓþÛÀàºÍKNN·ÖÀàËã·¨¿ÉÒÔ·ÖÎöÄÌÅ£Õ¾Á¢¡¢²½ÊýºÍÌÉÎÔ²¢½¨Ä£Ô¤²âõËÐеÄÄ£ÐÍ£¬×ÜÌå¼ì²â׼ȷÂÊ¿É´ïµ½87%,Ãô¸ÐÐÔΪ89.7%,ÌØÒìÐÔΪ72.5%[16].ɸѡ²»Í¬²¿Î»µÄÌØÕ÷²ÎÊýʱ£¬Ëλ³²¨µÈ[19]ÌáÈ¡ÄÌÅ£µÄÍ·²¿¡¢¾±²¿ÒÔ¼°Óë¾±Á¬½ÓµÄ±³²¿ÂÖÀªÏßÄâºÏÖ±ÏßбÂÊÊý¾Ý£¬KNN·ÖÀàËã·¨Ô¤²âõËÐеļì²âÕýÈ·ÂÊ¿É´ïµ½93.89%.µ«ÒÔÍȵÄÔ˶¯Ö¸±êΪõËÐÐÕï¶ÏµÄ±ê׼ʱ£¬¶ÌÆÚ¼ÇÒäÍøÂ磨long short-term memory, LSTM£©¡¢SVM¡¢KNNºÍDTËã·¨ÖУ¬»ùÓÚLSTMµÄõËÐмì²â׼ȷÂÊ×î¸ß£¬Îª98.57%[20].

¡¡¡¡2.3 Èé·¿Ñ×

¡¡¡¡MLÔ¤²âÈé·¿Ñ×ʱ£¬Ä̲úÁ¿¡¢¼·ÄÌʱ¼äºÍÈéÏÙÑ׵ȱêÖ¾ÎïÊý¾ÝÊÇÖ÷ÒªµÄÊý¾ÝÔ´£¬SVMºÍDTÊdz£ÓõÄËã·¨[21].RF¿ÉÒÔ½¨Á¢¶Ô»·¾³ÐÔÓ봫ȾÐÔÈéÏÙÑ×½øÐÐÇø±ðÕï¶ÏºÍ¶Ô¸ÉÄÌÆÚºÍÃÚÈéÆÚ¸ÐȾÇø±ðµÄÄ£ÐÍ[8].Éî¶Èѧϰ£¨deep learning, DL£©¡¢DT¡¢RF¡¢NB¡¢GBDT¡¢¹ãÒåÏßÐÔÄ£ÐͺÍÂß¼­»Ø¹éËã·¨Ô¤²âÑÇÁÙ´²ÈéÏÙÑ×ʱ£¬GBDTºÍDLÓнϸߵÄÔ¤²âÁéÃô¶È[9].Èç¹ûÊý¾ÝԴΪÈéÖ­µçµ¼ÂÊ£¬Ê¹ÓÃDT½¨Ä£Ô¤²âÈéÏÙÑ×µÄÌØÒìÐԿɸߴï99.2%[22].²»Í¬µÄDTËã·¨Ô¤²âÈéÏÙÑ×µÄ׼ȷÐÔÒ²ÓвîÒ죬DT¡¢Ê÷×®DF¡¢²¢ÐÐDTºÍRF 4ÖÖËã·¨ÖÐÖ»ÓÐRFÕï¶ÏÄÌÅ£ÈéÏÙÑ×µÄ׼ȷÂÊ¿É´ïµ½90%,²¢ÓÐÍûÔÚʵ¼ùÖÐʹÓÃ[23].´ËÍ⣬ͨ¹ýDTËã·¨¶Ô´ó³¦°£Ï£ÊϾú¸ÐȾÓÕ·¢ÈéÏÙÑ×ת¼×é·Ö²¼½øÐн¨Ä££¬¿Éɸѡ³ö´ó³¦°£Ï£ÊϾúÈéÏÙÑ׵ıêÖ¾Îï»ùÒò[24].

¡¡¡¡2.4 ÈÈÓ¦¼¤

¡¡¡¡ÄÌÅ£ÈÈÓ¦¼¤Óɼ«¶ËµÄÆøºò»·¾³Ôì³É£¬¼à²â»·¾³µÄÖ¸±êΪζÈ-ʪ¶ÈÖ¸Êý£¨temperature-humidity index, THI£©¡£µ«½öͨ¹ýTHI²¢²»ÄÜÖ±½Ó·´Ó³ÄÌÅ£»úÌåµÄÈÈÓ¦¼¤³Ì¶È[25],»¹ÐèÒªºôÎüƵÂÊ¡¢ºôÎüÆÀ·Ö¡¢ÌåΡ¢ÌÉÎÔÂÊ¡¢Õ¾Á¢Ê±¼äºÍÒûˮʱ¼äµÈÉúÀíÖ¸±ê[26].ÔÚËã·¨·½Ã棬»Ø¹é·ÖÎö£¬RFºÍANNʹÓÃ×î¶à¡£Gorczyca M TµÈ[10]±È½ÏÁËÏßÐԻع顢RF¡¢GBDTºÍANN½¨Ä£Ô¤²âÈÈÓ¦¼¤Ê±ÄÌÅ£µÄºôÎüƵÂÊ¡¢Æ¤·ôζȺÍÒõµÀζȵÄЧ¹û£¬·¢ÏÖRFºÍANNµÄÔ¤²âЧ¹û½ÏºÃ¡£Slob NµÈ[12]ʹÓÃRFºÍANNËã·¨½¨Ä£ºóÈÏΪÆøζÔÈÈÓ¦¼¤Ó°Ïì×î´ó¶ø·çËÙµÄ×÷ÓÃ×îС£¬µ«ÓÉÓÚÔ¤²â»áÊÜ»·¾³Ó°Ï죬׼ȷÐÔ¿ÉÄÜÄÑÒÔ±£Ö¤¡£ÎªÉ¸Ñ¡³ö×îÓÅÄ£ÐÍ£¬ÓÐÑо¿Õ߲ɼ¯ÁËTHI¡¢ºôÎüƵÂÊ¡¢ÎÔ´²Ê±¼ä¡¢ÌÉÎÔÆÚ¡¢×ܲ½Êý¡¢Á÷ÏÑ¡¢ºôÎüÆÀ·Ö¡¢ÒõÁ¹´¦»òÅçÁÜ´¦Í£Áôʱ¼ä¡¢Ìåϸ°ûÆÀ·Ö¡¢Íøθζȡ¢½à¾»¶ÈÆÀ·Ö¡¢Ä̲úÁ¿ÒÔ¼°ÈéÖ¬ÂÊ¡¢Èéµ°°×ÂʵÈÊý¾Ý£¬Ê¹ÓÃÂß¼­»Ø¹é¡¢¸ß˹NBºÍRF 3ÖÖMLËã·¨½øÐÐÔ¤²â²¢·¢ÏÖ׼ȷÂʶ¼ºÜ¸ß£¬Âß¼­»Ø¹éЧ¹û×îºÃ£¬Õâ±íÃ÷·ÇÏßÐԵķ½·¨Ð§¹ûºÃ[27].Brown-BrandlµÈ[28]ÔÚÀûÓúôÎüƵÂʺÍÌå±íζȽ¨Ä£Ô¤²âÈÈÓ¦¼¤Ê±·¢ÏÖ£¬Á½ÖֻعéÄ£ÐÍ¡¢Á½ÖÖÄ£ºýÍÆÀíϵͳºÍÒ»ÖÖÉñ¾­ÍøÂçÄ£ÐͶ¼¹ý¶ÈÔ¤²âµÍÇ¿¶ÈÈÈÓ¦¼¤»ò¸ß¹À¸ßÇ¿¶ÈÈÈÓ¦¼¤Ê±¶¯ÎïµÄÈÈÓ¦¼¤£¬ÆäÔ­Òò¿ÉÄÜÊÇÒòΪMLÄ£ÐÍÎÞ·¨ÌáÇ°Ô¤²âÌìÆø¡£

¡¡¡¡2.5 ´«È¾ÐÔ¼²²¡

¡¡¡¡MLÔÚ´«È¾²¡µÄÑо¿ÉÏ£¬Ö÷Òª¼¯ÖÐÔÚÀûÓô«È¾ÐÔ¼²²¡µÄÁ÷Ðв¡Ñ§ÌØÕ÷½øÐоÛÀà·ÖÎö£¬É¸Ñ¡³ö¹Ø¼ü·çÏÕÒò×Ó£¬¹¹½¨Ô¤²âºÍ¼à²â´«È¾ÐÔ¼²²¡µÄÄ£ÐÍ¡£±ÈÈ磬ÀûÓÃÂß¼­»Ø¹é·ÖÎöÄÌÅ£µÄÄêÁ䡢ƷÖֺͳöÉúʱ¼ä¶ÔÄÌÅ£¸ÐȾ¸±½áºËµÄÓ°Ï죬ÅжÏÕâЩÒòËØÊÇ·ñÊǸ±½áºËµÄ·çÏÕÒòËØ£¬½á¹û·¢ÏÖ³öÉúÈÕÆÚÓ븱½áºË¸ÐȾÓÐ×ÅÏÔÖøµÄÏà¹ØÐÔ£¨P<0.05£©[18].ÔÚËã·¨·½Ã棬RFºÍANNÈÔÊdz£Óõķ½·¨¡£ÓÐÑо¿ÕßÀûÓÃRFºÍÔöÇ¿»Ø¹éËã·¨£¬·ÖÎöÄÌţϸ¾ú»ùÒò×éÊý¾ÝºÍÑо¿Ï¸¾úÔÚ²»Í¬ÎïÖÖ¼äµÄ´«²¥ËÙ¶È[29].ͨ¹ýANN·ÖÎöÅ£ÄÌÖкìÍâ¹âÆ×Êý¾Ý²¢¹¹½¨ÄÌÅ£½áºË²¡Ô¤²âÄ£ÐÍʱ£¬Éî¶È¾í»ýÉñ¾­ÍøÂçµÄÔ¤²â׼ȷÐÔ¿ÉÒԸߴï95%[15].´ËÍ⣬Ҳ¿ÉÒÔÀûÓ÷ÖÀàºÍ»Ø¹éÊ÷Ä£ÐÍ£¬·ÖÎö¶à¸ö´«È¾²¡·çÏÕÒò×ÓÖ®¼äµÄ¹ØÁªÐԺ͸ßΣȺÌåµÄ·¢²¡¿ÉÄÜÐÔ£¬´Ó¶øÈ·¶¨¼²²¡µÄ»¼²¡ÂÊ¡¢Ê¼þ¼ì³öÂʺͲ¡Ê·[30].

¡¡¡¡3 Õ¹Íû

¡¡¡¡Í¨¹ý×ÛÊöMLÔÚÔ¤²âÄÌÅ£¼²²¡µÄÑо¿µÄÏà¹ØÎÄÏ×·¢ÏÖ£¬MLÔÚ´úлÐÔ¼²²¡ºÍõËÐз½ÃæµÄÓ¦ÓÃ×î¶à£¬¶øDTËã·¨ÒòÆä¼òµ¥Ò×ÐУ¬²¢ÇÒÔËÐÐËٶȿ죬ÔÚÄÌÅ£¼²²¡ÁìÓòÓ¦ÓÃ×î¹ã¡£Ä¿Ç°£¬ËäÈ»»ùÓÚMLµÄÄÌÅ£¼²²¡Ô¤²âÒѳÉÑо¿Èȵ㣬µ«ÓÉÓÚÊý¾ÝµÄÓÐÏÞÐԺ͸´ÔÓÐÔ£¬Ñо¿Ò²Óöµ½ºÜ¶àÀ§ÄÑ¡£ÔÚËã·¨·½Ã棬¾ÛÀà·ÖÎö¡¢Éî¶ÈѧϰºÍÇ¿»¯Ñ§Ï°µÈËã·¨µÄÓ¦Óý«»á³ÉΪÄÌÅ£¼²²¡¼à²âÁìÓòµÄÖ÷ÒªÑо¿·½Ïò¡£´ËÍ⣬ÀûÓÃMLºÍ»ùÒò×éѧÑо¿ÄÌÅ£¼²²¡Ò²ÊÇδÀ´¿É¹Ø×¢µÄ·½Ïò¡£

¡¡¡¡²Î¿¼ÎÄÏ×

¡¡¡¡[1] KONSTANTINOS L,PATRIZIA B,DIMITRIOS M,et al.Machine learning in agriculture:A review[J] .Sensors,2018,18£¨8£©£º2674.

¡¡¡¡[2] BORCHERS M R,CHANG Y M.PROUDFOOT K L.et al.Machine-leamning-based calving prediction from activity.lying. and ruminating behaviors in dairy cat

¡¡¡¡tle[J].J Dairy Sci,2017,100£¨7£©£º5664-5674.

¡¡¡¡[3] COCKBURN M.Review.Applic ation and prospective discussion of machine learning for the management of dairy farms[J].Animals ,2020,10£¨9£©£º1690.

¡¡¡¡[4] TANEJA M,BYABAZAIRE J,JALODIA N.et al.Machine learning based fog computing assisted data-driven approach for early lameness detection in dairycattle[J]. Comput Electron Agr,2020, 171:105286.

¡¡¡¡[5] XU W,KNEGSEL AT M V,VERVOORT J J M,et al.Prediction of metabolic status of dairy cows in early lactation with on-farm cow data and machine learning algorithms[J].J Dairy Sci,2019,102£¨11£©£º10186-10201.

¡¡¡¡[6] WAGNER N,ANTOINE V,MIALON M-M,et al.Machine learning to detect behavioural anomalies in dairy cows under subacute ruminal acidosis[J].ComputElectron Agr,2020, 170: 105233.

¡¡¡¡[7] TREMBLAY M,KAMMER M,LANGE H.et al.ldentifying poor metabolic adaptation during early lactation in dairy Cows using cluster analysis[J].J Dairy Sci,2018, 101£¨8£©£º7311-7321.

¡¡¡¡[8] HYDE R M,DOWN P M,BRADLEY A J,et al.Automated predic tion of mastitis infection patterns in dairy herds using machine learning[J]. Sci Reports,2020,10£¨1£©£º1-8.

¡¡¡¡[9] EBRAHIMI M,MOHAMMADI-DEHCHESHMEH M.EBRAHIMIE E.et al.Comprehensive analysis of mac hine learning models for prediction of sub-clinical mastitis:Deep Learning and Gradient-Boosted Trees outperformother models[J].Comput Biol Med,2019,114:103456.

¡¡¡¡[10] GORCZYCA M T,GEBREMEDHIN K G .Ranking of environmental heat stressors for dairy cows using machine learning algorithms[J] Comput Electron Agr,2020,168:105124.

¡¡¡¡[11] HEIDE E M M V D,VEERKAMP R F,PELT M L V,et al.Comparing regression,naive Bayes, and random forest methods in the prediction of inpidual survival to second lactation in Holstein cattle[J].J Dairy Sci,2019, 102£¨10£©£º9409-9421.

¡¡¡¡[12] SLOB N,CATAL C,KASSAHUN A.Applic ation of machine learning to improve dairy farm management:A systematic literature review[J].Prev Vet Med,2021,187:105237.

¡¡¡¡[13] PROBO M,PASCOTTINI O B,LEBL ANC S,et al.Association between metabolic diseases and the culling risk of high-yielding dairy cows in a transition management facility using survival and decision tree analysis[J].J Dairy Sci,2018, 101£¨10£©£º9419-9429.

¡¡¡¡[14] EHRET A,HOCHSTUHL D,KRATTENMACHER N,et al.Use of genomic and metabolic information as well as milk performance records for prediction ofsubclinical ketosis risk via artificial neural networks[J]J Dairy Sci,2015, 98£¨1£©£º322-329.

¡¡¡¡[15] DENHOLM S J,BRAND W,MITCHELL A P,et al.Predicting bovine tuberculosis status of dairy cows from mid-infrared spectral data of milk using deep learning[J].J Dairy Sci,2020, 103£¨10£©£º9355-9367.

¡¡¡¡[16] BYABAZAIRE J,OLARIU C,TANEJA M,et al.Lameness detection as a service:applic ation of machine learning to an internet of cattl[C]/2019 16th IEEEAnnual Consumer Communications & Networking Conference£¨CCNC£©¡£IEEE,2019:1-6.

¡¡¡¡[17] VAN HOEIJ R,KOK A.BRUCKMAIER R,et al.Relationship between metabolic status and behavior in dairy cows in week 4 of lactation[J].Animal,2019,13£¨3£©£º640-648.

¡¡¡¡[18] ZARE Y,SHOOK G E,COLLINS M T,et al.Evidence of birth seasonality and clustering of Mycobacterium avium subspecies paratuberculosis infection inUS dairy herds[J].Prev Vet Med,2013,112£¨3-4£©£º276-284.

¡¡¡¡[19]Ëλ³²¨£¬½ª²¨£¬ÎâÙ»£¬µÈ»ùÓÚÍ·¾±²¿ÂÖÀªÄâºÏÖ±ÏßбÂÊÌØÕ÷µÄÄÌÅ£õËÐмì²â·½·¨[J].Å©Òµ¹¤³Ìѧ±¨£¬ 2018,34£¨15£©£º190-199.[20] WU D,WU Q,YIN X,et al.Lameness detection of dairy cows based on the YOLOV3 deep learning algorithm and a relative step size characteristic vector

¡¡¡¡[J]. Biosyst Eng.2020,189:150-163.

¡¡¡¡[21] MAMMADOVA N,KESKIN 1.Application of the support vector machine to predict subclinical mastitis in dairy cattle[J] .Sci World J,2013,2013:603897.

¡¡¡¡[22] KAMPHUIS C,MOLLENHORST H,FEELDERS A.et al.Decision-tree induction to detect clinical mastitis with automatic milking[J] Comput Electron Agr,2010,70£¨1£©£º60-68.

¡¡¡¡[23] EBRAHIMIE E, EBRAHIMI F,EBRAHIMI M,et al.Hierarchical pattern recognition in milking parameters predicts mastitis prevalence[J]. Comput Electron Agr,2018,147:6-11.

¡¡¡¡[24] SHARIFI S,PAKDEL A.EBRAHIMI M.et al.Integration of machine learning and meta-analysis identifies the transcriptomic bio-signature of mastitis disease in cattle[J].PLoS One,2018,13£¨2£©£ºe0191227.

¡¡¡¡[25] DIKMEN S,HANSEN P J.Is the temperature-humidity index the best indicator of heat stress in lactating dairy cows in a subtropical environment?[J].J Dairy Sci,2009,92£¨1£©£º109-116.

¡¡¡¡[26] GORCZYCA M T,MAIA M H F,CAMPOS M A S,et al.Machine learning algorithms to predict core ,skin,and hair-coat temperatures of piglets[J] Comput EIectron Agr,2018, 151:286-294.

¡¡¡¡[27] BECKER C A.AGHALARI A,MARUFUZZAMAN M,et al.Predicting dairy cattle heat stress using mac hine learning techniques[J]J Dairy Sci,2021,104£¨1£©£º501-524.

¡¡¡¡[28] BROWN-BRANDL T M,JONES D D,WOLDT W E. Evaluating modeling techniques for catle heat stress predic tion[J] Biosyst Eng,2005,91£¨4£©£º513-524.

¡¡¡¡[29] CRISPELL J,BENTON C,BALAZ D.et al.Combining genomics and epidemiology to analyse bi-directional transmission of Mycobacterium bovis in a multi-host system[J].eLife,2019,8:e45833.

¡¡¡¡[30] ROMERO M P,CHANG Y-M,BRUNTON L A,et al.Decision tree machine learning applied to bovine tuberculosis risk factors to aid disease control decision making[J].Prev Vet Med,2020,175:104860.

·µ»Ø±¾ÆªÂÛÎĵ¼º½
×÷Õßµ¥Î»£ºÎ÷±±Å©ÁֿƼ¼´óѧÐÅÏ¢¹¤³ÌѧԺ Î÷±±Å©ÁֿƼ¼´óѧ¶¯ÎïҽѧԺ
Ô­Îijö´¦£º·ëåû,¸ßÖ¾Ìì,Ö£ì¿çÍ,ÑîÖÙÌÎ,¶­Ç¿.»úÆ÷ѧϰÔÚÄÌÅ£ÁÙ´²¼²²¡Ô¤²âÖеÄÓ¦ÓÃ[J].¶¯Îïҽѧ½øÕ¹,2021,42(06):115-119.
Ïà¹ØÄÚÈÝÍƼö
Ïà¹Ø±êÇ©£º
·µ»Ø£º¼ÆËã»úÓ¦Óü¼ÊõÂÛÎÄ