欢迎来到天天文库
浏览记录
ID:49404396
大小:3.26 MB
页数:86页
时间:2020-03-01
《增量学习在电子鼻智能烘烤系统中的应用研究.pdf》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库。
1、໔䠅ᆜҖ൞⭫ᆆ啱Ჰ㜳✎✚㌱㔕ѣⲺᓊ⭞⹊ガ䠃ᒼཝᆜ⺋༡ᆜփ䇰ᮽδщѐᆜփεᆜ⭕ဉφᕖڛሲᮏᐾφ⭦䙘ᱛᮏᦾᆜփ㊱ࡡφᐛぁ⺋༡δ⭫ᆆф䙐ؗᐛぁ亼ตε䠃ᒼཝᆜ䙐ؗᐛぁᆜ䲘ӂOжޣᒪᴾStudyontheIncrementalLearningAppliedintheIntelligentCuringSystemsBasedonElectronicNoseAThesisSubmittedtoChongqingUniversityinPartialFulfillmentoftheRequirementforProfessionalDegreeByZhangJianSupervisedBy˖Prof
2、.TianFengchunSpecialty˖ME(FieldofElectronicsandCommunicationEngineering)CollegeofCommunicationEngineeringofChongqingUniversity,Chongqing,ChinaApril2016ѝ᮷᪈㾱㾷໎䟿ᆖҐѪᵪಘᆖҐⲴањ䟽㾱⹄ウᯩੁˈ䲿⵰བྷᮠᦞᰦԓⲴࡠᶕˈ⭡Ҿަн䎆ޘ䜘ṧᵜˈਚ࡙⭘઼؍ᆈṧᵜ⢩ᖱǃᾲᘥᡆԓ㺘ᙗⲴṧᵜˈ㜭ᴹ᭸ᇎ⧠ሩṧᵜⲴ㓖߿ᒦӾᯠṧᵜѝᆖҐࡠᯠⲴ⸕䇶ˈ൘䀓ߣᆈ䲀ࡦ઼䟽༽ᆖҐ䰞仈кᴹ⤜⢩Ո࣯ˈ㻛㿶Ѫབྷᮠᦞ䰞仈Ⲵа䟽㾱фᴹ᭸Ⲵ䀓ߣ࣎⌅DŽ≄ણᱟ⢙䍘Ⲵ䟽㾱⢩ᖱ
3、ѻаˈ⭥ᆀ啫ᢰᵟᱟаสҾԯ⭏௵㿹⨶Ⲵᯠޤᰐᦏ≄ણỰ⍻ᢰᵟˈ∄ҾՐ㔏ⲴॆᆖǃᆖỰ⍻ᯩ⌅ˈަާᴹᘛ䙏ǃ߶⺞ǃᇒ㿲ˈׯᦧⲴՈ⛩ˈᱟԯ⭏௵㿹亶ฏѝⲴ⹄ウ✝⛩DŽ䲿⵰“ᐕъ4.0”ᰦԓⲴࡠᶕˈᲪ㜭ॆᐢᡀѪᵚᶕᐕъⲴѫ㾱ਁኅ䎻࣯DŽⴞࡽᡁഭሩᐕъ✈✔ѝⲴᲪ㜭ॆ⹄ウ䖳ቁˈ✈✔䗷〻ѫ㾱䶐ӪⲴ㓿傼઼㠚ࣘॆᵪಘ㔃ਸᇎ⧠ˈަоⵏ↓ⲴᰐӪᲪ㜭㠚ࣘ✈✔ᐞ䖳䘌DŽᵜ᮷䪸ሩ✏㥹ᐕъⲴ䟽㾱⧟㢲——✏ਦ✈✔ѝⲴᲪ㜭ॆǃ㠚ࣘॆ≤ᒣվ䰞仈ˈ⹄ウҶᇎ⧠✏ਦ㠚ࣘᲪ㜭✈✔Ⲵᯩ⌅DŽ・䏣Ҿ㠚ѫ⹄ਁⲴ✈✔ؑॆ᧗ᒣਠˈ࡙⭘⭥ᆀ啫ǃᩴؑ࠶≤઼ۿമˈણ≄Ⲵਦ✏Ҷ䳶᭦ᦏᰐˈ༷䇮䳶䟷ㅹᵪۿˈ᧒䇘Ҷ✏ਦ⢩ᖱ൘✈✔䗷〻ѝⲴਈॆ㿴ᖻˈᒦ㔃ਸ᭦䳶
4、ࡠⲴ✈✔ᐸ䈳᧗✈✔ᐕ㢪ᴢ㓯ؑˈᔪ・Ҷ⭡✏ਦ⢩ᖱؑ亴⍻✈✔ᐕ㢪৲ᮠⲴᵪಘᆖҐ㖁㔌⁑රDŽ਼ᰦˈ䪸ሩн਼૱䍘Ⲵ✏ਦ൘✈✔ѝⲴ༽ᵲᙗˈ⹄ウҶᓄሩ䈕ᛵߥлⲴ༴⨶ᯩ⌅——໎䟿ᆖҐDŽ⭡Ҿ✏ਦ✈✔亴⍻㖁㔌ᱟа㊫എᖂ亴⍻㖁㔌ˈᵜ᮷ѫ㾱⹄ウҶสҾ᭟ᤱੁ䟿എᖂᵪⲴ໎䟿ᆖҐᯩ⌅DŽ൘࠶᷀Ҷ⧠ᴹสҾ᭟ᤱੁ䟿Ⲵ໎䟿㇇⌅㕪⛩৺✈✔ᆖҐṧᵜ໎䟿⢩⛩ਾˈᨀࠪҶ⭘㠚㓴㓷ㄎҹ㖁㔌(SOINN)㧧ਆṧᵜⲴᤃᢁ㢲⛩ؑˈ࡙⭘࣐ᵳ᭟ᤱੁ䟿എᖂᵪሩᮠᦞ䘋㹼࣐ᵳ䇝㓳Ⲵᯩ⌅DŽ䈕ᯩ⌅㜭䖳ྭൠ؍⮉ᮠᦞⲴ⢩⛩ᒦ߿ᕡа㡜สҾ᭟ᤱੁ䟿Ⲵ໎䟿⁑ර൘໎䟿ᆖҐѝⲴ㕪⛩DŽ䪸ሩSOINN㖁㔌൘໎䟿ᢩᮠᦞᾲᘥ┲〫ᛵߥлᆈ൘Ⲵ䰞仈ᨀࠪҶᓄⲴ᭩䘋༴⨶ᯩ⌅
5、DŽ㇇⌅൘Ӫᐕsin࠭ᮠᮠᦞ䳶ǃUCIᮠᦞ䳶৺ᇎ⍻ᮠᦞ䳶л⍻䈅ᒦ傼䇱Ҷ᭸᷌DŽᴰਾ・䏣ҾᐢᴹⲴᲪ㜭✈✔᧗ࡦᒣਠˈ䇮䇑Ҷ൘ᇎ䱵ᓄ⭘ѝⲴ✏ਦᲪ㜭✈✔㌫㔏ᐕ⍱〻ˈѪᇎ⧠Ც㜭ॆ✈✔ᨀҶُ䢤о৲㘳DŽީ䭤䈃˖໎䟿ᆖҐˈᲪ㜭᧗ࡦˈ✈✔㌫㔏ˈ࣐ᵳ᭟ᤱੁ䟿എᖂᵪˈ⭥ᆀ啫I㤡᮷᪈㾱ABSTRACTAsoneoftheimportantresearchdirectioninmachinelearning,comingwiththeageofbigdata,incrementallearningisconsideredasanimportantandeffectivesolutiontosolvetheprobl
6、emoflargedatasinceitdoesnotneedtorelyonallsamples,onlyuseandpreservethecharacteristics,conceptslearnedformerorrepresentativesamples,caneffectivelyachievethesamplereductionandlearningnewknowledge,hasauniqueadvantageindealingwithmemorylimitationandrepeatedlearning.Smellisoneoftheimportantcharacteris
7、ticsofthematerial,E-nose(electronicnose)technologyisanewnondestructivesmelldetectiontechnologybasedonthebionicolfactionprinciple.Comparetothetraditionalchemicalandopticaldetectionmethod,thee-nosemethodisrapid,acc
此文档下载收益归作者所有