WEKA浣跨敤鏁欑▼
鐩綍
1. 绠€浠?br />2. 鏁版嵁鏍煎紡
3.鏁版嵁鍑嗗
4. 鍏宠仈瑙勫垯锛堣喘鐗╃鍒嗘瀽锛?br />5. 鍒嗙被涓庡洖褰?br />6. 鑱氱被鍒嗘瀽
1. 绠€浠?/strong>
WEKA鐨勫叏鍚嶆槸鎬€鍗℃墭鏅鸿兘鍒嗘瀽鐜锛圵aikato Environment for Knowledge Analysis锛夛紝瀹冪殑婧愪唬鐮佸彲閫氳繃http://www.cs.waikato.ac.nz/ml/weka寰楀埌銆傚悓鏃秝eka涔熸槸鏂拌タ鍏扮殑涓€绉嶉笩鍚嶏紝鑰學EKA鐨勪富瑕佸紑鍙戣€呮潵鑷柊瑗垮叞銆?br />
WEKA浣滀负涓€涓叕寮€鐨勬暟鎹寲鎺樺伐浣滃钩鍙帮紝闆嗗悎浜嗗ぇ閲忚兘鎵挎媴鏁版嵁鎸栨帢浠诲姟鐨勬満鍣ㄥ涔犵畻娉曪紝鍖呮嫭瀵规暟鎹繘琛岄澶勭悊锛屽垎绫伙紝鍥炲綊銆佽仛绫汇€佸叧鑱旇鍒欎互鍙婂湪鏂扮殑浜や簰寮忕晫闈笂鐨勫彲瑙嗗寲銆?br />濡傛灉鎯宠嚜宸卞疄鐜版暟鎹寲鎺樼畻娉曠殑璇濓紝鍙互鐪嬩竴鐪媤eka鐨勬帴鍙ｆ枃妗ｃ€傚湪weka涓泦鎴愯嚜宸辩殑绠楁硶鐢氳嚦鍊熼壌瀹冪殑鏂规硶鑷繁瀹炵幇鍙鍖栧伐鍏峰苟涓嶆槸浠跺緢鍥伴毦鐨勪簨鎯呫€?nbsp;
2005骞?鏈堬紝鍦ㄧ11灞夾CM SIGKDD鍥介檯浼氳涓婏紝鎬€鍗℃墭澶у鐨刉eka灏忕粍鑽ｈ幏浜嗘暟鎹寲鎺樺拰鐭ヨ瘑鎺㈢储棰嗗煙鐨勬渶楂樻湇鍔″锛學eka绯荤粺寰楀埌浜嗗箍娉涚殑璁ゅ彲锛岃瑾変负鏁版嵁鎸栨帢鍜屾満鍣ㄥ涔犲巻鍙蹭笂鐨勯噷绋嬬锛屾槸鐜颁粖鏈€瀹屽鐨勬暟鎹寲鎺樺伐鍏蜂箣涓€锛堝凡鏈?1骞寸殑鍙戝睍鍘嗗彶锛夈€俉eka鐨勬瘡鏈堜笅杞芥鏁板凡瓒呰繃涓囨銆?br />
--鏁寸悊鑷?a href="http://www.china-pub.com/computers/common/info.asp?id=29304" target="_blank">http://www.china-pub.com/computers/common/info.asp?id=29304
2. 鏁版嵁鏍煎紡
宸у闅句负鏃犵背涔嬬倞銆傞鍏堟垜浠潵鐪嬬湅WEKA鎵€鐢ㄧ殑鏁版嵁搴旀槸浠€涔堟牱鐨勬牸寮忋€?nbsp;
璺熷緢澶氱數瀛愯〃鏍兼垨鏁版嵁鍒嗘瀽杞欢涓€鏍凤紝WEKA鎵€澶勭悊鐨勬暟鎹泦鏄浘1閭ｆ牱鐨勪竴涓簩缁寸殑琛ㄦ牸銆?nbsp;
![鎶€鏈垎浜? src=]()
鍥? 鏂扮獥鍙ｆ墦寮€
杩欓噷鎴戜滑瑕佷粙缁嶄竴涓媁EKA涓殑鏈銆傝〃鏍奸噷鐨勪竴涓í琛岀О浣滀竴涓疄渚嬶紙Instance锛夛紝鐩稿綋浜庣粺璁″涓殑涓€涓牱鏈紝鎴栬€呮暟鎹簱涓殑涓€鏉¤褰曘€傜珫琛岀О浣滀竴涓睘鎬э紙Attrbute锛夛紝鐩稿綋浜庣粺璁″涓殑涓€涓彉閲忥紝鎴栬€呮暟鎹簱涓殑涓€涓瓧娈点€傝繖鏍蜂竴涓〃鏍硷紝鎴栬€呭彨鏁版嵁闆嗭紝鍦╓EKA鐪嬫潵锛屽憟鐜颁簡灞炴€т箣闂寸殑涓€绉嶅叧绯?Relation)銆傚浘1涓竴鍏辨湁14涓疄渚嬶紝5涓睘鎬э紝鍏崇郴鍚嶇О涓?ldquo;weather”銆?br />
WEKA瀛樺偍鏁版嵁鐨勬牸寮忔槸ARFF锛圓ttribute-Relation File Format锛夋枃浠讹紝杩欐槸涓€绉岮SCII鏂囨湰鏂囦欢銆傚浘1鎵€绀虹殑浜岀淮琛ㄦ牸瀛樺偍鍦ㄥ涓嬬殑ARFF鏂囦欢涓€傝繖涔熷氨鏄疻EKA鑷甫鐨?ldquo;weather.arff”鏂囦欢锛屽湪WEKA瀹夎鐩綍鐨?ldquo;data”瀛愮洰褰曚笅鍙互鎵惧埌銆?/p>
![鎶€鏈垎浜? src=]()
闇€瑕佹敞鎰忕殑鏄紝鍦╓indows璁颁簨鏈墦寮€杩欎釜鏂囦欢鏃讹紝鍙兘浼氬洜涓哄洖杞︾瀹氫箟涓嶄竴鑷磋€屽鑷村垎琛屼笉姝ｅ父銆傛帹鑽愪娇鐢?a name="OLE_LINK37">UltraEdit杩欐牱鐨勫瓧绗︾紪杈戣蒋浠跺療鐪婣RFF鏂囦欢鐨勫唴瀹广€?br />
涓嬮潰鎴戜滑鏉ュ杩欎釜鏂囦欢鐨勫唴瀹硅繘琛岃鏄庛€?nbsp;
璇嗗埆ARFF鏂囦欢鐨勯噸瑕佷緷鎹槸鍒嗚锛屽洜姝や笉鑳藉湪杩欑鏂囦欢閲岄殢鎰忕殑鏂銆傜┖琛岋紙鎴栧叏鏄┖鏍肩殑琛岋級灏嗚蹇界暐銆?nbsp;
浠?ldquo;%”寮€濮嬬殑琛屾槸娉ㄩ噴锛學EKA灏嗗拷鐣ヨ繖浜涜銆傚鏋滀綘鐪嬪埌鐨?ldquo;weather.arff”鏂囦欢澶氫簡鎴栧皯浜嗕簺“%”寮€濮嬬殑琛岋紝鏄病鏈夊奖鍝嶇殑銆?nbsp;
闄ゅ幓娉ㄩ噴鍚庯紝鏁翠釜ARFF鏂囦欢鍙互鍒嗕负涓や釜閮ㄥ垎銆傜涓€閮ㄥ垎缁欏嚭浜嗗ご淇℃伅锛圚ead information锛夛紝鍖呮嫭浜嗗鍏崇郴鐨勫０鏄庡拰瀵瑰睘鎬х殑澹版槑銆傜浜岄儴鍒嗙粰鍑轰簡鏁版嵁淇℃伅锛圖ata information锛夛紝鍗虫暟鎹泦涓粰鍑虹殑鏁版嵁銆備粠“@data”鏍囪寮€濮嬶紝鍚庨潰鐨勫氨鏄暟鎹俊鎭簡銆?br />
鍏崇郴澹版槑
鍏崇郴鍚嶇О鍦ˋRFF鏂囦欢鐨勭涓€涓湁鏁堣鏉ュ畾涔夛紝鏍煎紡涓?nbsp;
@relation <relation-name>
<relation-name>鏄竴涓瓧绗︿覆銆傚鏋滆繖涓瓧绗︿覆鍖呭惈绌烘牸锛屽畠蹇呴』鍔犱笂寮曞彿锛堟寚鑻辨枃鏍囩偣鐨勫崟寮曞彿鎴栧弻寮曞彿锛夈€?nbsp;
灞炴€у０鏄?nbsp;
灞炴€у０鏄庣敤涓€鍒椾互“@attribute”寮€澶寸殑璇彞琛ㄧず銆傛暟鎹泦涓殑姣忎竴涓睘鎬ч兘鏈夊畠瀵瑰簲鐨?ldquo;@attribute”璇彞锛屾潵瀹氫箟瀹冪殑灞炴€у悕绉板拰鏁版嵁绫诲瀷銆?nbsp;
杩欎簺澹版槑璇彞鐨勯『搴忓緢閲嶈銆傞鍏堝畠琛ㄦ槑浜嗚椤瑰睘鎬у湪鏁版嵁閮ㄥ垎鐨勪綅缃€備緥濡傦紝“humidity”鏄涓変釜琚０鏄庣殑灞炴€э紝杩欒鏄庢暟鎹儴鍒嗛偅浜涜閫楀彿鍒嗗紑鐨勫垪涓紝绗笁鍒楁暟鎹?85 90 86 96 ... 鏄浉搴旂殑“humidity”鍊笺€傚叾娆★紝鏈€鍚庝竴涓０鏄庣殑灞炴€ц绉颁綔class灞炴€э紝鍦ㄥ垎绫绘垨鍥炲綊浠诲姟涓紝瀹冩槸榛樿鐨勭洰鏍囧彉閲忋€?/strong>
灞炴€у０鏄庣殑鏍煎紡涓?nbsp;
@attribute <attribute-name> <datatype>
鍏朵腑<attribute-name>鏄繀椤讳互瀛楁瘝寮€澶寸殑瀛楃涓层€傚拰鍏崇郴鍚嶇О涓€鏍凤紝濡傛灉杩欎釜瀛楃涓插寘鍚┖鏍硷紝瀹冨繀椤诲姞涓婂紩鍙枫€?nbsp;
WEKA鏀寔鐨?lt;datatype>鏈夊洓绉嶏紝鍒嗗埆鏄?nbsp;
numeric-------------------------鏁板€煎瀷
<nominal-specification>-----鍒嗙被锛坣ominal锛夊瀷
string----------------------------瀛楃涓插瀷
date [<date-format>]--------鏃ユ湡鍜屾椂闂村瀷
鍏朵腑<nominal-specification> 鍜?lt;date-format> 灏嗗湪涓嬮潰璇存槑銆傝繕鍙互浣跨敤涓や釜绫诲瀷“integer”鍜?ldquo;real”锛屼絾鏄疻EKA鎶婂畠浠兘褰撲綔“numeric”鐪嬪緟銆傛敞鎰?ldquo;integer”锛?ldquo;real”锛?ldquo;numeric”锛?ldquo;date”锛?ldquo;string”杩欎簺鍏抽敭瀛楁槸鍖哄垎澶у皬鍐欑殑锛岃€?ldquo;relation”“attribute ”鍜?ldquo;date”鍒欎笉鍖哄垎銆?br />
鏁板€煎睘鎬?nbsp;
鏁板€煎瀷灞炴€у彲浠ユ槸鏁存暟鎴栬€呭疄鏁帮紝浣哤EKA鎶婂畠浠兘褰撲綔瀹炴暟鐪嬪緟銆?nbsp;
鍒嗙被灞炴€?nbsp;
鍒嗙被灞炴€х敱<nominal-specification>鍒楀嚭涓€绯诲垪鍙兘鐨勭被鍒悕绉板苟鏀惧湪鑺辨嫭鍙蜂腑锛歿<nominal-name1>, <nominal-name2>, <nominal-name3>, ...} 銆傛暟鎹泦涓灞炴€х殑鍊煎彧鑳芥槸鍏朵腑涓€绉嶇被鍒€?br />渚嬪濡備笅鐨勫睘鎬у０鏄庤鏄?ldquo;outlook”灞炴€ф湁涓夌绫诲埆锛?ldquo;sunny”锛?ldquo; overcast”鍜?ldquo;rainy”銆傝€屾暟鎹泦涓瘡涓疄渚嬪搴旂殑“outlook”鍊煎繀鏄繖涓夎€呬箣涓€銆?br />@attribute outlook {sunny, overcast, rainy}
濡傛灉绫诲埆鍚嶇О甯︽湁绌烘牸锛屼粛闇€瑕佸皢涔嬫斁鍏ュ紩鍙蜂腑銆?nbsp;
瀛楃涓插睘鎬?nbsp;
瀛楃涓插睘鎬т腑鍙互鍖呭惈浠绘剰鐨勬枃鏈€傝繖绉嶇被鍨嬬殑灞炴€у湪鏂囨湰鎸栨帢涓潪甯告湁鐢ㄣ€?nbsp;
绀轰緥锛?nbsp;
@ATTRIBUTE LCC string
鏃ユ湡鍜屾椂闂村睘鎬?nbsp;
鏃ユ湡鍜屾椂闂村睘鎬х粺涓€鐢?ldquo;date”绫诲瀷琛ㄧず锛屽畠鐨勬牸寮忔槸
@attribute <name> date [<date-format>]
鍏朵腑<name>鏄繖涓睘鎬х殑鍚嶇О锛?lt;date-format>鏄竴涓瓧绗︿覆锛屾潵瑙勫畾璇ユ€庢牱瑙ｆ瀽鍜屾樉绀烘棩鏈熸垨鏃堕棿鐨勬牸寮忥紝榛樿鐨勫瓧绗︿覆鏄疘SO-8601鎵€缁欑殑鏃ユ湡鏃堕棿缁勫悎鏍煎紡“yyyy-MM-ddTHH:mm:ss”銆?br />鏁版嵁淇℃伅閮ㄥ垎琛ㄨ揪鏃ユ湡鐨勫瓧绗︿覆蹇呴』绗﹀悎澹版槑涓瀹氱殑鏍煎紡瑕佹眰锛堜笅鏂囨湁渚嬪瓙锛夈€?nbsp;
鏁版嵁淇℃伅
鏁版嵁淇℃伅涓?ldquo;@data”鏍囪鐙崰涓€琛岋紝鍓╀笅鐨勬槸鍚勪釜瀹炰緥鐨勬暟鎹€?nbsp;
姣忎釜瀹炰緥鍗犱竴琛屻€傚疄渚嬬殑鍚勫睘鎬у€肩敤閫楀彿“,”闅斿紑銆?strong>濡傛灉鏌愪釜灞炴€х殑鍊兼槸缂哄け鍊硷紙missing value锛夛紝鐢ㄩ棶鍙?ldquo;?”琛ㄧず锛屼笖杩欎釜闂彿涓嶈兘鐪佺暐銆備緥濡傦細
@data
sunny,85,85,FALSE,no
?,78,90,?,yes
瀛楃涓插睘鎬у拰鍒嗙被灞炴€х殑鍊兼槸鍖哄垎澶у皬鍐欑殑銆傝嫢鍊间腑鍚湁绌烘牸锛屽繀椤昏寮曞彿鎷捣鏉ャ€備緥濡傦細
@relation LCCvsLCSH
@attribute LCC string
@attribute LCSH string
@data
AG5, 鈥楨ncyclopedias and dictionaries.;Twentieth century.鈥?nbsp;
AS262, 鈥楽cience -- Soviet Union -- History.鈥?nbsp;
鏃ユ湡灞炴€х殑鍊煎繀椤讳笌灞炴€у０鏄庝腑缁欏畾鐨勭浉涓€鑷淬€備緥濡傦細
@RELATION Timestamps
@ATTRIBUTE timestamp DATE "yyyy-MM-dd HH:mm:ss"
@DATA
"2001-04-03 12:12:12"
"2001-05-03 12:59:55"
绋€鐤忔暟鎹?nbsp;
鏈夌殑鏃跺€欐暟鎹泦涓惈鏈夊ぇ閲忕殑0鍊硷紙姣斿璐墿绡垎鏋愶級锛岃繖涓椂鍊欑敤绋€鐤忔牸寮忕殑鏁版嵁瀛樿串鏇村姞鐪佺┖闂淬€?nbsp;
绋€鐤忔牸寮忔槸閽堝鏁版嵁淇℃伅涓煇涓疄渚嬬殑琛ㄧず鑰岃█锛屼笉闇€瑕佷慨鏀笰RFF鏂囦欢鐨勫叾瀹冮儴鍒嗐€傜湅濡備笅鐨勬暟鎹細
@data
0, X, 0, Y, "class A"
0, 0, W, 0, "class B"
鐢ㄧ█鐤忔牸寮忚〃杈剧殑璇濆氨鏄?nbsp;
@data
{1 X, 3 Y, 4 "class A"}
{2 W, 4 "class B"}
姣忎釜瀹炰緥鐢ㄨ姳鎷彿鎷捣鏉ャ€傚疄渚嬩腑姣忎竴涓潪0鐨勫睘鎬у€肩敤<index> <绌烘牸> <value>琛ㄧず銆?lt;index>鏄睘鎬х殑搴忓彿锛屼粠0寮€濮嬭锛?lt;value>鏄睘鎬у€笺€傚睘鎬у€间箣闂翠粛鐢ㄩ€楀彿闅斿紑銆傝繖閲屾瘡涓疄渚嬬殑鏁板€煎繀椤绘寜灞炴€х殑椤哄簭鏉ュ啓锛屽 {1 X, 3 Y, 4 "class A"}锛屼笉鑳藉啓鎴恵3 Y, 1 X, 4 "class A"}銆?br />娉ㄦ剰鍦ㄧ█鐤忔牸寮忎腑娌℃湁娉ㄦ槑鐨勫睘鎬у€间笉鏄己澶卞€硷紝鑰屾槸0鍊笺€傝嫢瑕佽〃绀虹己澶卞€煎繀椤绘樉寮忕殑鐢ㄩ棶鍙疯〃绀哄嚭鏉ャ€?br />
Relational鍨嬪睘鎬?nbsp;
鍦╓EKA 3.5鐗堜腑澧炲姞浜嗕竴绉嶅睘鎬х被鍨嬪彨鍋歊elational锛屾湁浜嗚繖绉嶇被鍨嬫垜浠彲浠ュ儚鍏崇郴鍨嬫暟鎹簱閭ｆ牱澶勭悊澶氫釜缁村害浜嗐€備絾鏄繖绉嶇被鍨嬬洰鍓嶈繕涓嶈骞挎硾搴旂敤锛屾殏涓嶄綔浠嬬粛銆?br />
--鏁寸悊鑷?a href="http://www.cs.waikato.ac.nz/~ml/weka/arff.html" target="_blank">http://www.cs.waikato.ac.nz/~ml/weka/arff.html 鍜?a href="http://weka.sourceforge.net/wekadoc/index.php/en:ARFF_%283.5.3%29" target="_blank">http://weka.sourceforge.net/wekadoc/index.php/en:ARFF_%283.5.3%29
3.鏁版嵁鍑嗗
浣跨敤WEKA浣滄暟鎹寲鎺橈紝闈复鐨勭涓€涓棶棰樺線寰€鏄垜浠殑鏁版嵁涓嶆槸ARFF鏍煎紡鐨勩€傚垢濂斤紝WEKA杩樻彁渚涗簡瀵笴SV鏂囦欢鐨勬敮鎸侊紝鑰岃繖绉嶆牸寮忔槸琚緢澶氬叾浠栬蒋浠舵墍鏀寔鐨勩€傛澶栵紝WEKA杩樻彁渚涗簡閫氳繃JDBC璁块棶鏁版嵁搴撶殑鍔熻兘銆?br />鍦ㄨ繖涓€鑺傞噷锛屾垜浠厛浠xcel鍜孧atlab涓轰緥锛岃鏄庡浣曡幏寰桟SV鏂囦欢銆傜劧鍚庢垜浠皢鐭ラ亾CSV鏂囦欢濡備綍杞寲鎴怉RFF鏂囦欢锛屾瘯绔熷悗鑰呮墠鏄疻EKA鏀寔寰楁渶濂界殑鏂囦欢鏍煎紡銆傞潰瀵逛竴涓狝RFF鏂囦欢锛屾垜浠粛鏈変竴浜涢澶勭悊瑕佸仛锛屾墠鑳借繘琛屾寲鎺樹换鍔°€?br />
.* -> .csv
鎴戜滑缁欏嚭涓€涓狢SV鏂囦欢鐨勪緥瀛愶紙bank-data.csv锛夈€傜敤UltraEdit鎵撳紑瀹冨彲浠ョ湅鍒帮紝杩欑鏍煎紡涔熸槸涓€绉嶉€楀彿鍒嗗壊鏁版嵁鐨勬枃鏈枃浠?鍌ㄥ瓨浜嗕竴涓簩缁磋〃鏍笺€?br />
Excel鐨刋LS鏂囦欢鍙互璁╁涓簩缁磋〃鏍兼斁鍒颁笉鍚岀殑宸ヤ綔琛紙Sheet锛変腑锛屾垜浠彧鑳芥妸姣忎釜宸ヤ綔琛ㄥ瓨鎴愪笉鍚岀殑CSV鏂囦欢銆傛墦寮€涓€涓猉LS鏂囦欢骞跺垏鎹㈠埌闇€瑕佽浆鎹㈢殑宸ヤ綔琛紝鍙﹀瓨涓篊SV绫诲瀷锛岀偣“纭畾”銆?ldquo;鏄?rdquo;蹇界暐鎻愮ず鍗冲彲瀹屾垚鎿嶄綔銆?br />
鍦∕atlab涓殑浜岀淮琛ㄦ牸鏄竴涓煩闃碉紝鎴戜滑閫氳繃杩欐潯鍛戒护鎶婁竴涓煩闃靛瓨鎴怌SV鏍煎紡銆?nbsp;
csvwrite(鈥榝ilename鈥?matrixname)
闇€瑕佹敞鎰忕殑鏄紝Matllab缁欏嚭鐨凜SV鏂囦欢寰€寰€娌℃湁灞炴€у悕锛圗xcel缁欏嚭鐨勪篃鏈夊彲鑳芥病鏈夛級銆?strong>鑰學EKA蹇呴』浠嶤SV鏂囦欢鐨勭涓€琛岃鍙栧睘鎬у悕锛屽惁鍒欏氨浼氭妸绗竴琛岀殑鍚勫睘鎬у€艰鎴愬彉閲忓悕銆?/strong>鍥犳鎴戜滑瀵逛簬Matllab缁欏嚭鐨凜SV鏂囦欢闇€瑕佺敤UltraEdit鎵撳紑锛屾墜宸ユ坊鍔犱竴琛屽睘鎬у悕銆傛敞鎰忓睘鎬у悕鐨勪釜鏁拌璺熸暟鎹睘鎬х殑涓暟涓€鑷达紝浠嶇敤閫楀彿闅斿紑銆?br />
.csv -> .arff
灏咰SV杞崲涓篈RFF鏈€杩呮嵎鐨勫姙娉曟槸浣跨敤WEKA鎵€甯︾殑鍛戒护琛屽伐鍏枫€?nbsp;
杩愯WEKA鐨勪富绋嬪簭锛屽嚭鐜癎UI鍚庡彲浠ョ偣鍑讳笅鏂规寜閽繘鍏ョ浉搴旂殑妯″潡銆傛垜浠偣鍑昏繘鍏?ldquo;Simple CLI”妯″潡鎻愪緵鐨勫懡浠よ鍔熻兘銆傚湪鏂扮獥鍙ｇ殑鏈€涓嬫柟锛堜笂鏂规槸涓嶈兘鍐欏瓧鐨勶級杈撳叆妗嗗啓涓?br />java weka.core.converters.CSVLoader filename.csv > filename.arff
鍗冲彲瀹屾垚杞崲銆?nbsp;
鍦╓EKA 3.5涓彁渚涗簡涓€涓?ldquo;Arff Viewer”妯″潡锛屾垜浠彲浠ョ敤瀹冩墦寮€涓€涓狢SV鏂囦欢灏嗚繘琛屾祻瑙堬紝鐒跺悗鍙﹀瓨涓篈RFF鏂囦欢銆?nbsp;
杩涘叆“Exploer”妯″潡锛屼粠涓婃柟鐨勬寜閽腑鎵撳紑CSV鏂囦欢鐒跺悗鍙﹀瓨涓篈RFF鏂囦欢浜﹀彲銆?nbsp;
“Exploer”鐣岄潰
鎴戜滑搴旇娉ㄦ剰鍒帮紝“Exploer”杩樻彁渚涗簡寰堝鍔熻兘锛屽疄闄呬笂鍙互璇磋繖鏄疻EKA浣跨敤鏈€澶氱殑妯″潡銆傜幇鍦ㄦ垜浠厛鏉ョ啛鎮夊畠鐨勭晫闈紝鐒跺悗鍒╃敤瀹冨鏁版嵁杩涜棰勫鐞嗐€?nbsp;
![鎶€鏈垎浜? src=]()
鍥? 鏂扮獥鍙ｆ墦寮€
鍥?鏄剧ず鐨勬槸浣跨敤3.5鐗?Exploer"鎵撳紑"bank-data.csv"鐨勬儏鍐点€傛垜浠牴鎹笉鍚岀殑鍔熻兘鎶婅繖涓晫闈㈠垎鎴?涓尯鍩熴€?nbsp;
鍖哄煙1鐨勫嚑涓€夐」鍗℃槸鐢ㄦ潵鍒囨崲涓嶅悓鐨勬寲鎺樹换鍔￠潰鏉裤€傝繖涓€鑺傜敤鍒扮殑鍙湁“Preprocess”锛屽叾浠栭潰鏉跨殑鍔熻兘灏嗗湪浠ュ悗浠嬬粛銆?nbsp;
鍖哄煙2鏄竴浜涘父鐢ㄦ寜閽€傚寘鎷墦寮€鏁版嵁锛屼繚瀛樺強缂栬緫鍔熻兘銆傛垜浠湪杩欓噷鎶?bank-data.csv"鍙﹀瓨涓?bank-data.arff"銆?nbsp;
鍦ㄥ尯鍩?涓?ldquo;Choose”鏌愪釜“Filter”锛屽彲浠ュ疄鐜扮瓫閫夋暟鎹垨鑰呭鏁版嵁杩涜鏌愮鍙樻崲銆傛暟鎹澶勭悊涓昏灏卞埄鐢ㄥ畠鏉ュ疄鐜般€?nbsp;
鍖哄煙4灞曠ず浜嗘暟鎹泦鐨勪竴浜涘熀鏈儏鍐点€?nbsp;
鍖哄煙5涓垪鍑轰簡鏁版嵁闆嗙殑鎵€鏈夊睘鎬с€傚嬀閫変竴浜涘睘鎬у苟“Remove”灏卞彲浠ュ垹闄ゅ畠浠紝鍒犻櫎鍚庤繕鍙互鍒╃敤鍖哄煙2鐨?ldquo;Undo”鎸夐挳鎵惧洖銆傚尯鍩?涓婃柟鐨勪竴鎺掓寜閽槸鐢ㄦ潵瀹炵幇蹇€熷嬀閫夌殑銆?br />鍦ㄥ尯鍩?涓€変腑鏌愪釜灞炴€э紝鍒欏尯鍩?涓湁鍏充簬杩欎釜灞炴€х殑鎽樿銆傛敞鎰忓浜庢暟鍊煎睘鎬у拰鍒嗙被灞炴€э紝鎽樿鐨勬柟寮忔槸涓嶄竴鏍风殑銆傚浘涓樉绀虹殑鏄鏁板€煎睘鎬?ldquo;income”鐨勬憳瑕併€?nbsp;
鍖哄煙7鏄尯鍩?涓€変腑灞炴€х殑鐩存柟鍥俱€傝嫢鏁版嵁闆嗙殑鏈€鍚庝竴涓睘鎬э紙鎴戜滑璇磋繃杩欐槸鍒嗙被鎴栧洖褰掍换鍔＄殑榛樿鐩爣鍙橀噺锛夋槸鍒嗙被鍙橀噺锛堣繖閲岀殑“pep”姝ｅソ鏄級锛岀洿鏂瑰浘涓殑姣忎釜闀挎柟褰㈠氨浼氭寜鐓ц鍙橀噺鐨勬瘮渚嬪垎鎴愪笉鍚岄鑹茬殑娈点€傝鎯虫崲涓垎娈电殑渚濇嵁锛屽湪鍖哄煙7涓婃柟鐨勪笅鎷夋涓€変釜涓嶅悓鐨勫垎绫诲睘鎬у氨鍙互浜嗐€備笅鎷夋閲岄€変笂“No Class”鎴栬€呬竴涓暟鍊煎睘鎬т細鍙樻垚榛戠櫧鐨勭洿鏂瑰浘銆?br />鍖哄煙8鏄姸鎬佹爮锛屽彲浠ユ煡鐪婰og浠ュ垽鏂槸鍚︽湁閿欍€傚彸杈圭殑weka楦熷湪鍔ㄧ殑璇濊鏄嶹EKA姝ｅ湪鎵ц鎸栨帢浠诲姟銆傚彸閿偣鍑荤姸鎬佹爮杩樺彲浠ユ墽琛孞AVA鍐呭瓨鐨勫瀮鍦惧洖鏀躲€?nbsp;
棰勫鐞?nbsp;
bank-data鏁版嵁鍚勫睘鎬х殑鍚箟濡備笅锛?nbsp;
id a unique identification number
age age of customer in years (numeric)
sex MALE / FEMALE
region inner_city/rural/suburban/town
income income of customer (numeric)
married is the customer married (YES/NO)
children number of children (numeric)
car does the customer own a car (YES/NO)
save_acct does the customer have a saving account (YES/NO)
current_acct does the customer have a current account (YES/NO)
mortgage does the customer have a mortgage (YES/NO)
pep did the customer buy a PEP (Personal Equity Plan) after the last mailing (YES/NO)
閫氬父瀵逛簬鏁版嵁鎸栨帢浠诲姟鏉ヨ锛孖D杩欐牱鐨勪俊鎭槸鏃犵敤鐨勶紝鎴戜滑灏嗕箣鍒犻櫎銆傚湪鍖哄煙5鍕鹃€夊睘鎬?ldquo;id”锛屽苟鐐瑰嚮“Remove”銆傚皢鏂扮殑鏁版嵁闆嗕繚瀛樹竴娆★紝骞剁敤UltraEdit鎵撳紑杩欎釜ARFF鏂囦欢銆傛垜浠彂鐜帮紝鍦ㄥ睘鎬у０鏄庨儴鍒嗭紝WEKA宸茬粡涓烘瘡涓睘鎬ч€夊ソ浜嗗悎閫傜殑绫诲瀷銆?br />
鎴戜滑鐭ラ亾锛屾湁浜涚畻娉曪紝鍙兘澶勭悊鎵€鏈夌殑灞炴€ч兘鏄垎绫诲瀷鐨勬儏鍐点€傝繖鏃跺€欐垜浠氨闇€瑕佸鏁板€煎瀷鐨勫睘鎬ц繘琛岀鏁ｅ寲銆傚湪杩欎釜鏁版嵁闆嗕腑鏈?涓彉閲忔槸鏁板€煎瀷鐨勶紝鍒嗗埆鏄?ldquo;age”锛?ldquo;income”鍜?ldquo;children”銆?br />鍏朵腑“children”鍙湁4涓彇鍊硷細0锛?锛?锛?銆傝繖鏃舵垜浠湪UltraEdit涓洿鎺ヤ慨鏀笰RFF鏂囦欢锛屾妸
@attribute children numeric
鏀逛负
@attribute children {0,1,2,3}
灏卞彲浠ヤ簡銆?nbsp;
鍦?ldquo;Explorer”涓噸鏂版墦寮€“bank-data.arff”锛岀湅鐪嬮€変腑“children”灞炴€у悗锛屽尯鍩?閭ｉ噷鏄剧ず鐨?ldquo;Type”鏄笉鏄彉鎴?ldquo;Nominal”浜嗭紵
“age”鍜?ldquo;income”鐨勭鏁ｅ寲鎴戜滑闇€瑕佸€熷姪WEKA涓悕涓?ldquo;Discretize”鐨凢ilter鏉ュ畬鎴愩€傚湪鍖哄煙2涓偣“Choose”锛屽嚭鐜颁竴妫?ldquo;Filter鏍?rdquo;锛岄€愮骇鎵惧埌“weka.filters.unsupervised.attribute.Discretize”锛岀偣鍑汇€傝嫢鏃犳硶鍏抽棴杩欎釜鏍戯紝鍦ㄦ爲涔嬪鐨勫湴鏂圭偣鍑?ldquo;Explorer”闈㈡澘鍗冲彲銆?br />鐜板湪“Choose”鏃佽竟鐨勬枃鏈搴旇鏄剧ず“Discretize -B 10 -M -0.1 -R first-last”銆?鐐瑰嚮杩欎釜鏂囨湰妗嗕細寮瑰嚭鏂扮獥鍙ｄ互淇敼绂绘暎鍖栫殑鍙傛暟銆?br />鎴戜滑涓嶆墦绠楀鎵€鏈夌殑灞炴€х鏁ｅ寲锛屽彧鏄拡瀵瑰绗?涓拰绗?涓睘鎬э紙瑙佸尯鍩?灞炴€у悕宸﹁竟鐨勬暟瀛楋級锛屾晠鎶奱ttributeIndices鍙宠竟鏀规垚“1,4”銆傝鍒掓妸杩欎袱涓睘鎬ч兘鍒嗘垚3娈碉紝浜庢槸鎶?ldquo;bins”鏀规垚“3”銆傚叾瀹冩閲屼笉鐢ㄦ洿鏀癸紝鍏充簬瀹冧滑鐨勬剰鎬濆彲浠ョ偣“More”鏌ョ湅銆傜偣“OK”鍥炲埌“Explorer”锛屽彲浠ョ湅鍒?ldquo;age”鍜?ldquo;income”宸茬粡琚鏁ｅ寲鎴愬垎绫诲瀷鐨勫睘鎬с€傝嫢鎯虫斁寮冪鏁ｅ寲鍙互鐐瑰尯鍩?鐨?ldquo;Undo”銆?br />濡傛灉瀵?ldquo;"(-inf-34.333333]"”杩欐牱鏅︽订鐨勬爣璇嗕笉婊★紝鎴戜滑鍙互鐢║ltraEdit鎵撳紑淇濆瓨鍚庣殑ARFF鏂囦欢锛屾妸鎵€鏈夌殑“鈥榎鈥?-inf-34.333333]\鈥樷€?rdquo;鏇挎崲鎴?ldquo;0_34”銆傚叾瀹冩爣璇嗗仛绫讳技鍦版墜鍔ㄦ浛鎹€?br />
缁忚繃涓婅堪鎿嶄綔寰楀埌鐨勬暟鎹泦鎴戜滑淇濆瓨涓?a href="http://maya.cs.depaul.edu/~classes/ect584/WEKA/data/bank-data-final.arff" target="_blank">bank-data-final.arff銆?br />
----鏁寸悊鑷?a href="http://maya.cs.depaul.edu/~classes/ect584/WEKA/preprocess.html" target="_blank">http://maya.cs.depaul.edu/~classes/ect584/WEKA/preprocess.html
4. 鍏宠仈瑙勫垯锛堣喘鐗╃鍒嗘瀽锛?/strong>
娉ㄦ剰锛氱洰鍓嶏紝WEKA鐨勫叧鑱旇鍒欏垎鏋愬姛鑳戒粎鑳界敤鏉ヤ綔绀鸿寖锛屼笉閫傚悎鐢ㄦ潵鎸栨帢澶у瀷鏁版嵁闆嗐€?nbsp;
鎴戜滑鎵撶畻瀵瑰墠闈㈢殑“bank-data”鏁版嵁浣滃叧鑱旇鍒欑殑鍒嗘瀽銆傜敤“Explorer”鎵撳紑“bank-data-final.arff”鍚庯紝鍒囨崲鍒?ldquo;Associate”閫夐」鍗°€傞粯璁ゅ叧鑱旇鍒欏垎鏋愭槸鐢ˋpriori绠楁硶锛屾垜浠氨鐢ㄨ繖涓畻娉曪紝浣嗘槸鐐?ldquo;Choose”鍙宠竟鐨勬枃鏈淇敼榛樿鐨勫弬鏁帮紝寮瑰嚭鐨勭獥鍙ｄ腑鐐?ldquo;More”鍙互鐪嬪埌鍚勫弬鏁扮殑璇存槑銆?br />
鑳屾櫙鐭ヨ瘑
棣栧厛鎴戜滑鏉ユ俯涔犱竴涓婣priori鐨勬湁鍏崇煡璇嗐€傚浜庝竴鏉″叧鑱旇鍒橪->R锛屾垜浠父鐢ㄦ敮鎸佸害锛圫upport锛夊拰缃俊搴︼紙Confidence锛夋潵琛￠噺瀹冪殑閲嶈鎬с€傝鍒欑殑鏀寔搴︽槸鐢ㄦ潵浼拌鍦ㄤ竴涓喘鐗╃涓悓鏃惰瀵熷埌L鍜孯鐨勬鐜嘝(L,R)锛岃€岃鍒欑殑缃俊搴︽槸浼拌璐墿鏍忎腑鍑虹幇浜哃鏃朵篃鍑轰細鐜癛鐨勬潯浠舵鐜嘝(R|L)銆傚叧鑱旇鍒欑殑鐩爣涓€鑸槸浜х敓鏀寔搴﹀拰缃俊搴﹂兘杈冮珮鐨勮鍒欍€?br />鏈夊嚑涓被浼肩殑搴﹂噺浠ｆ浛缃俊搴︽潵琛￠噺瑙勫垯鐨勫叧鑱旂▼搴︼紝瀹冧滑鍒嗗埆鏄?nbsp;
Lift锛堟彁鍗囧害锛燂級锛?nbsp;P(L,R)/(P(L)P(R))
Lift=1鏃惰〃绀篖鍜孯鐙珛銆傝繖涓暟瓒婂ぇ锛岃秺琛ㄦ槑L鍜孯瀛樺湪鍦ㄤ竴涓喘鐗╃涓笉鏄伓鐒剁幇璞°€?nbsp;
Leverage锛堜笉鐭ラ亾鎬庝箞缈昏瘧锛夛細P(L,R)-P(L)P(R)
瀹冨拰Lift鐨勫惈涔夊樊涓嶅銆侺everage=0鏃禠鍜孯鐙珛锛孡everage瓒婂ぇL鍜孯鐨勫叧绯昏秺瀵嗗垏銆?nbsp;
Conviction锛堟洿涓嶇煡閬撹瘧浜嗭級锛歅(L)P(!R)/P(L,!R) 锛?R琛ㄧずR娌℃湁鍙戠敓锛?br />Conviction涔熸槸鐢ㄦ潵琛￠噺L鍜孯鐨勭嫭绔嬫€с€備粠瀹冨拰lift鐨勫叧绯伙紙瀵筊鍙栧弽锛屼唬鍏ift鍏紡鍚庢眰鍊掓暟锛夊彲浠ョ湅鍑猴紝鎴戜滑涔熷笇鏈涜繖涓€艰秺澶ц秺濂姐€?nbsp;
鍊煎緱娉ㄦ剰鐨勬槸锛岀敤Lift鍜孡everage浣滄爣鍑嗘椂锛孡鍜孯鏄绉扮殑锛孋onfidence鍜孋onviction鍒欎笉鐒躲€?nbsp;
鍙傛暟璁剧疆
鐜板湪鎴戜滑璁″垝鎸栨帢鍑烘敮鎸佸害鍦?0%鍒?00%涔嬮棿锛屽苟涓攍ift鍊艰秴杩?.5涓攍ift鍊兼帓鍦ㄥ墠100浣嶇殑閭ｄ簺鍏宠仈瑙勫垯銆傛垜浠妸“lowerBoundMinSupport”鍜?ldquo;upperBoundMinSupport”鍒嗗埆璁句负0.1鍜?锛?ldquo;metricType”璁句负lift锛?ldquo;minMetric”璁句负1.5锛?ldquo;numRules”璁句负100銆傚叾浠栭€夐」淇濇寔榛樿鍗冲彲銆?ldquo;OK” 涔嬪悗鍦?ldquo;Explorer”涓偣鍑?ldquo;Start”寮€濮嬭繍琛岀畻娉曪紝鍦ㄥ彸杈圭獥鍙ｆ樉绀烘暟鎹泦鎽樿鍜屾寲鎺樼粨鏋溿€?br />
涓嬮潰鏄寲鎺樺嚭鏉ョ殑lift鎺掑墠5鐨勮鍒欍€?nbsp;
Best rules found:
1. age=52_max save_act=YES current_act=YES 113 ==> income=43759_max 61 conf:(0.54) < lift:(4.05)> lev:(0.0 [45] conv:(1.85)
2. income=43759_max 80 ==> age=52_max save_act=YES current_act=YES 61 conf:(0.76) < lift:(4.05)> lev:(0.0 [45] conv:(3.25)
3. income=43759_max current_act=YES 63 ==> age=52_max save_act=YES 61 conf:(0.97) < lift:(3.85)> lev:(0.0 [45] conv:(15.72)
4. age=52_max save_act=YES 151 ==> income=43759_max current_act=YES 61 conf:(0.4) < lift:(3.85)> lev:(0.0 [45] conv:(1.49)
5. age=52_max save_act=YES 151 ==> income=43759_max 76 conf:(0.5) < lift:(3.77)> lev:(0.09) [55] conv:(1.72)
瀵逛簬鎸栨帢鍑虹殑姣忔潯瑙勫垯锛學EKA鍒楀嚭浜嗗畠浠叧鑱旂▼搴︾殑鍥涢」鎸囨爣銆?nbsp;
鍛戒护琛屾柟寮?nbsp;
鎴戜滑涔熷彲浠ュ埄鐢ㄥ懡浠よ鏉ュ畬鎴愭寲鎺樹换鍔★紝鍦?ldquo;Simlpe CLI”妯″潡涓緭鍏ュ涓嬫牸寮忕殑鍛戒护锛?nbsp;
java weka.associations.Apriori options -t directory-path\bank-data-final.arff
鍗冲彲瀹屾垚Apriori绠楁硶銆傛敞鎰忥紝“-t”鍙傛暟鍚庣殑鏂囦欢璺緞涓笉鑳藉惈鏈夌┖鏍笺€?nbsp;
鍦ㄥ墠闈㈡垜浠娇鐢ㄧ殑option涓?nbsp;
-N 100 -T 1 -C 1.5 -D 0.05 -U 1.0 -M 0.1 -S -1.0 鍛戒护琛屼腑浣跨敤杩欎簺鍙傛暟寰楀埌鐨勭粨鏋滃拰鍓嶉潰鍒╃敤GUI寰楀埌鐨勪竴鏍枫€?nbsp;
鎴戜滑杩樺彲浠ュ姞涓?ldquo;- I”鍙傛暟锛屽緱鍒颁笉鍚岄」鏁扮殑棰戠箒椤归泦銆傛垜鐢ㄧ殑鍛戒护濡備笅锛?nbsp;
java weka.associations.Apriori -N 100 -T 1 -C 1.5 -D 0.05 -U 1.0 -M 0.1 -S -1.0 -I -t d:\weka\bank-data-final.arff
鎸栨帢缁撴灉鍦ㄤ笂鏂规樉绀猴紝搴旀槸杩欎釜鏂囦欢鐨勬牱瀛愩€?br />
----鏁寸悊鑷?a href="http://maya.cs.depaul.edu/~classes/ect584/WEKA/associate.html" target="_blank">http://maya.cs.depaul.edu/~classes/ect584/WEKA/associate.html
5. 鍒嗙被涓庡洖褰?/strong>
鑳屾櫙鐭ヨ瘑
WEKA鎶婂垎绫?Classification)鍜屽洖褰?Regression)閮芥斁鍦?ldquo;Classify”閫夐」鍗′腑锛岃繖鏄湁鍘熷洜鐨勩€?nbsp;
鍦ㄨ繖涓や釜浠诲姟涓紝閮芥湁涓€涓洰鏍囧睘鎬э紙杈撳嚭鍙橀噺锛夈€傛垜浠笇鏈涙牴鎹竴涓牱鏈?WEKA涓О浣滃疄渚?鐨勪竴缁勭壒寰侊紙杈撳叆鍙橀噺锛夛紝瀵圭洰鏍囪繘琛岄娴嬨€備负浜嗗疄鐜拌繖涓€鐩殑锛屾垜浠渶瑕佹湁涓€涓缁冩暟鎹泦锛岃繖涓暟鎹泦涓瘡涓疄渚嬬殑杈撳叆鍜岃緭鍑洪兘鏄凡鐭ョ殑銆傝瀵熻缁冮泦涓殑瀹炰緥锛屽彲浠ュ缓绔嬭捣棰勬祴鐨勬ā鍨嬨€傛湁浜嗚繖涓ā鍨嬶紝鎴戜滑灏卞彲浠ユ柊鐨勮緭鍑烘湭鐭ョ殑瀹炰緥杩涜棰勬祴浜嗐€傝 閲忔ā鍨嬬殑濂藉潖灏卞湪浜庨娴嬬殑鍑嗙‘绋嬪害銆?br />鍦╓EKA涓紝寰呴娴嬬殑鐩爣锛堣緭鍑猴級琚О浣淐lass灞炴€э紝杩欏簲璇ユ槸鏉ヨ嚜鍒嗙被浠诲姟鐨?ldquo;绫?rdquo;銆?strong>涓€鑸殑锛岃嫢Class灞炴€ф槸鍒嗙被鍨嬫椂鎴戜滑鐨勪换鍔℃墠鍙垎绫伙紝Class灞炴€ф槸鏁板€煎瀷鏃舵垜浠殑浠诲姟鍙洖褰掋€?br />
閫夋嫨绠楁硶
杩欎竴鑺備腑锛屾垜浠娇鐢–4.5鍐崇瓥鏍戠畻娉曞bank-data寤虹珛璧峰垎绫绘ā鍨嬨€?nbsp;
鎴戜滑鏉ョ湅鍘熸潵鐨?ldquo;bank-data.csv”鏂囦欢銆?ldquo;ID”灞炴€ц偗瀹氭槸涓嶉渶瑕佺殑銆傜敱浜嶤4.5绠楁硶鍙互澶勭悊鏁板€煎瀷鐨勫睘鎬э紝鎴戜滑涓嶇敤鍍忓墠闈㈢敤鍏宠仈瑙勫垯閭ｆ牱鎶婃瘡涓彉閲忛兘绂绘暎鍖栨垚鍒嗙被鍨嬨€傚敖绠″姝わ紝鎴戜滑杩樻槸鎶?ldquo;Children”灞炴€ц浆鎹㈡垚鍒嗙被鍨嬬殑涓や釜鍊?ldquo;YES”鍜?ldquo;NO”銆傚彟澶栵紝鎴戜滑鐨勮缁冮泦浠呭彇鍘熸潵鏁版嵁闆嗗疄渚嬬殑涓€鍗婏紱鑰屼粠鍙﹀涓€鍗婁腑鎶藉嚭鑻ュ共鏉′綔涓哄緟棰勬祴鐨勫疄渚嬶紝瀹冧滑鐨?ldquo;pep”灞炴€ч兘璁句负缂哄け鍊笺€傜粡杩囦簡杩欎簺澶勭悊鐨勮缁冮泦鏁版嵁鍦?a href="http://maya.cs.depaul.edu/~classes/ect584/WEKA/classify/bank.arff" target="_blank">杩欓噷涓嬭浇锛涘緟棰勬祴闆嗘暟鎹湪杩欓噷涓嬭浇銆?br />
鎴戜滑鐢?ldquo;Explorer”鎵撳紑璁粌闆?ldquo;bank.arff”锛岃瀵熶竴涓嬪畠鏄笉鏄寜鐓у墠闈㈢殑瑕佹眰澶勭悊濂戒簡銆傚垏鎹㈠埌“Classify”閫夐」鍗★紝鐐瑰嚮“Choose”鎸夐挳鍚庡彲浠ョ湅鍒板緢澶氬垎绫绘垨鑰呭洖褰掔殑绠楁硶鍒嗛棬鍒被鐨勫垪鍦ㄤ竴涓爲鍨嬫閲屻€?.5鐗堢殑WEKA涓紝鏍戝瀷妗嗕笅鏂规湁涓€涓?ldquo;Filter...”鎸夐挳锛岀偣鍑诲彲浠ユ牴鎹暟鎹泦鐨勭壒鎬ц繃婊ゆ帀涓嶅悎閫傜殑绠楁硶銆傛垜浠暟鎹泦鐨勮緭鍏ュ睘鎬т腑鏈?ldquo;Binary”鍨嬶紙鍗冲彧鏈変袱涓被鐨勫垎绫诲瀷锛夊拰鏁板€煎瀷鐨勫睘鎬э紝鑰孋lass鍙橀噺鏄?ldquo;Binary”鐨勶紱浜庢槸鎴戜滑鍕鹃€?ldquo;Binary attributes”“Numeric attributes”鍜?ldquo;Binary class”銆傜偣“OK”鍚庡洖鍒版爲褰㈠浘锛屽彲浠ュ彂鐜颁竴浜涚畻娉曞悕绉板彉绾簡锛岃鏄庡畠浠笉鑳界敤銆傞€夋嫨“trees”涓嬬殑“J48”锛岃繖灏辨槸鎴戜滑闇€瑕佺殑C4.5绠楁硶锛岃繕濂藉畠娌℃湁鍙樼孩銆?br />鐐瑰嚮“Choose”鍙宠竟鐨勬枃鏈锛屽脊鍑烘柊绐楀彛涓鸿绠楁硶璁剧疆鍚勭鍙傛暟銆傜偣“More”鏌ョ湅鍙傛暟璇存槑锛岀偣“Capabilities”鏄煡鐪嬬畻娉曢€傜敤鑼冨洿銆傝繖閲屾垜浠妸鍙傛暟淇濇寔榛樿銆?br />鐜板湪鏉ョ湅宸︿腑鐨?ldquo;Test Option”銆傛垜浠病鏈変笓闂ㄨ缃楠屾暟鎹泦锛屼负浜嗕繚璇佺敓鎴愮殑妯″瀷鐨勫噯纭€ц€屼笉鑷充簬鍑虹幇杩囨嫙鍚堬紙overfitting锛夌殑鐜拌薄锛屾垜浠湁蹇呰閲囩敤10鎶樹氦鍙夐獙璇侊紙10-fold cross validation锛夋潵閫夋嫨鍜岃瘎浼版ā鍨嬨€傝嫢涓嶆槑鐧戒氦鍙夐獙璇佺殑鍚箟鍙互Google涓€涓嬨€?br />
寤烘ā缁撴灉
OK锛岄€変笂“Cross-validation”骞跺湪“Folds”妗嗗～涓?ldquo;10”銆傜偣“Start”鎸夐挳寮€濮嬭绠楁硶鐢熸垚鍐崇瓥鏍戞ā鍨嬨€傚緢蹇紝鐢ㄦ枃鏈〃绀虹殑涓€妫靛喅绛栨爲锛屼互鍙婂杩欎釜鍐崇瓥鏍戠殑璇樊鍒嗘瀽绛夌瓑缁撴灉鍑虹幇鍦ㄥ彸杈圭殑“Classifier output”涓€傚悓鏃跺乏涓嬬殑“Results list”鍑虹幇浜嗕竴涓」鐩樉绀哄垰鎵嶇殑鏃堕棿鍜岀畻娉曞悕绉般€傚鏋滄崲涓€涓ā鍨嬫垨鑰呮崲涓弬鏁帮紝閲嶆柊“Start”涓€娆★紝鍒?ldquo;Results list”鍙堜細澶氬嚭涓€椤广€?br />
鎴戜滑鐪嬪埌“J48”绠楁硶浜ゅ弶楠岃瘉鐨勭粨鏋滀箣涓€涓?nbsp;
Correctly Classified Instances 206 68.6667 %
涔熷氨鏄杩欎釜妯″瀷鐨勫噯纭害鍙湁69%宸﹀彸銆備篃璁告垜浠渶瑕佸鍘熷睘鎬ц繘琛屽鐞嗭紝鎴栬€呬慨鏀圭畻娉曠殑鍙傛暟鏉ユ彁楂樺噯纭害銆備絾杩欓噷鎴戜滑涓嶇瀹冿紝缁х画鐢ㄨ繖涓ā鍨嬨€?nbsp;
鍙抽敭鐐瑰嚮“Results list”鍒氭墠鍑虹幇鐨勯偅涓€椤癸紝寮瑰嚭鑿滃崟涓€夋嫨“Visualize tree”锛屾柊绐楀彛閲屽彲浠ョ湅鍒板浘褰㈡ā寮忕殑鍐崇瓥鏍戙€傚缓璁妸杩欎釜鏂扮獥鍙ｆ渶澶у寲锛岀劧鍚庣偣鍙抽敭锛岄€?ldquo;Fit to screen”锛屽彲浠ユ妸杩欎釜鏍戠湅娓呮浜涖€傜湅瀹屽悗鎴浘鎴栬€呭叧鎺?/strong>
杩欓噷鎴戜滑瑙ｉ噴涓€涓?ldquo;Confusion Matrix”鐨勫惈涔夈€?nbsp;
=== Confusion Matrix ===
a b <-- classified as
74 64 | a = YES
30 132 | b = NO
杩欎釜鐭╅樀鏄锛屽師鏈?ldquo;pep”鏄?ldquo;YES”鐨勫疄渚嬶紝鏈?4涓姝ｇ‘鐨勯娴嬩负“YES”锛屾湁64涓敊璇殑棰勬祴鎴愪簡“NO”锛涘師鏈?ldquo;pep”鏄?ldquo;NO”鐨勫疄渚嬶紝鏈?0涓閿欒鐨勯娴嬩负“YES”锛屾湁132涓纭殑棰勬祴鎴愪簡“NO”銆?4+64+30+132 = 300鏄疄渚嬫€绘暟锛岃€?74+132)/300 = 0.68667姝ｅソ鏄纭垎绫荤殑瀹炰緥鎵€鍗犳瘮渚嬨€傝繖涓煩闃靛瑙掔嚎涓婄殑鏁板瓧瓒婂ぇ锛岃鏄庨娴嬪緱瓒婂ソ銆?br />
妯″瀷搴旂敤
鐜板湪鎴戜滑瑕佺敤鐢熸垚鐨勬ā鍨嬪閭ｄ簺寰呴娴嬬殑鏁版嵁闆嗚繘琛岄娴嬩簡銆傛敞鎰忓緟棰勬祴鏁版嵁闆嗗拰璁粌鐢ㄦ暟鎹泦鍚勪釜灞炴€х殑璁剧疆蹇呴』鏄竴鑷寸殑銆傚嵆浣夸綘娌℃湁寰呴娴嬫暟鎹泦鐨凜lass灞炴€х殑鍊硷紝浣犱篃瑕佹坊鍔犺繖涓睘鎬э紝鍙互灏嗚灞炴€у湪鍚勫疄渚嬩笂鐨勫€煎潎璁炬垚缂哄け鍊笺€?br />鍦?ldquo;Test Opion”涓€夋嫨“Supplied test set”锛屽苟涓?ldquo;Set”鎴愪綘瑕佸簲鐢ㄦā鍨嬬殑鏁版嵁闆嗭紝杩欓噷鏄?ldquo;bank-new.arff”鏂囦欢銆?br />鐜板湪锛屽彸閿偣鍑?ldquo;Result list”涓垰浜х敓鐨勯偅涓€椤癸紝閫夋嫨“Re-evaluate model on current test set”銆傚彸杈规樉绀虹粨鏋滅殑鍖哄煙涓細澧炲姞涓€浜涘唴瀹癸紝鍛婅瘔浣犺妯″瀷搴旂敤鍦ㄨ繖涓暟鎹泦涓婅〃鐜板皢濡備綍銆傚鏋滀綘鐨凜lass灞炴€ч兘鏄簺缂哄け鍊硷紝閭ｈ繖浜涘唴瀹规槸鏃犳剰涔夌殑锛屾垜浠叧娉ㄧ殑鏄ā鍨嬪湪鏂版暟鎹泦涓婄殑棰勬祴鍊笺€?br />鐜板湪鐐瑰嚮鍙抽敭鑿滃崟涓殑“Visualize classifier errors”锛屽皢寮瑰嚭涓€涓柊绐楀彛鏄剧ず涓€浜涙湁鍏抽娴嬭宸殑鏁ｇ偣鍥俱€傜偣鍑昏繖涓柊绐楀彛涓殑“Save”鎸夐挳锛屼繚瀛樹竴涓狝rff鏂囦欢銆傛墦寮€杩欎釜鏂囦欢鍙互鐪嬪埌鍦ㄥ€掓暟绗簩涓綅缃浜嗕竴涓睘鎬э紙predictedpep锛夛紝杩欎釜灞炴€т笂鐨勫€煎氨鏄ā鍨嬪姣忎釜瀹炰緥鐨勯娴嬪€笺€?br />
浣跨敤鍛戒护琛岋紙鎺ㄨ崘锛?nbsp;
铏界劧浣跨敤鍥惧舰鐣岄潰鏌ョ湅缁撴灉鍜岃缃弬鏁板緢鏂逛究锛屼絾鏄渶鐩存帴鏈€鐏垫椿鐨勫缓妯″強搴旂敤鐨勫姙娉曚粛鏄娇鐢ㄥ懡浠よ銆?nbsp;
鎵撳紑“Simple CLI”妯″潡锛屽儚涓婇潰閭ｆ牱浣跨敤“J48”绠楁硶鐨勫懡浠ゆ牸寮忎负锛?nbsp;
java weka.classifiers.trees.J48 -C 0.25 -M 2 -t directory-path\bank.arff -d directory-path \bank.model
鍏朵腑鍙傛暟“ -C 0.25”鍜?ldquo;-M 2”鏄拰鍥惧舰鐣岄潰涓墍璁剧殑涓€鏍风殑銆?ldquo;-t ”鍚庨潰璺熺潃鐨勬槸璁粌鏁版嵁闆嗙殑瀹屾暣璺緞锛堝寘鎷洰褰曞拰鏂囦欢鍚嶏級锛?ldquo;-d ”鍚庨潰璺熺潃鐨勬槸淇濆瓨妯″瀷鐨勫畬鏁磋矾寰勩€傛敞鎰忥紒杩欓噷鎴戜滑鍙互鎶婃ā鍨嬩繚瀛樹笅鏉ャ€?br />杈撳叆涓婅堪鍛戒护鍚庯紝鎵€寰楀埌鏍戞ā鍨嬪拰璇樊鍒嗘瀽浼氬湪“Simple CLI”涓婃柟鏄剧ず锛屽彲浠ュ鍒朵笅鏉ヤ繚瀛樺湪鏂囨湰鏂囦欢閲屻€傝宸槸鎶婃ā鍨嬪簲鐢ㄥ埌璁粌闆嗕笂缁欏嚭鐨勩€?br />鎶婅繖涓ā鍨嬪簲鐢ㄥ埌“bank-new.arff”鎵€鐢ㄥ懡浠ょ殑鏍煎紡涓猴細
java weka.classifiers.trees.J48 -p 9 -l directory-path\bank.model -T directory-path \bank-new.arff
鍏朵腑“-p 9”璇寸殑鏄ā鍨嬩腑鐨勫緟棰勬祴灞炴€х殑鐪熷疄鍊煎瓨鍦ㄧ9涓紙涔熷氨鏄?ldquo;pep”锛夊睘鎬т腑锛岃繖閲屽畠浠叏閮ㄦ湭鐭ュ洜姝ゅ叏閮ㄧ敤缂哄け鍊间唬鏇裤€?ldquo;-l”鍚庨潰鏄ā鍨嬬殑瀹屾暣璺緞銆?ldquo;-T”鍚庨潰鏄緟棰勬祴鏁版嵁闆嗙殑瀹屾暣璺緞銆?br />杈撳叆涓婅堪鍛戒护鍚庯紝鍦?ldquo;Simple CLI”涓婃柟浼氭湁杩欐牱涓€浜涚粨鏋滐細
0 YES 0.75 ?
1 NO 0.7272727272727273 ?
2 YES 0.95 ?
3 YES 0.8813559322033898 ?
4 NO 0.8421052631578947 ?
...
杩欓噷鐨勭涓€鍒楀氨鏄垜浠彁鍒拌繃鐨?ldquo;Instance_number”锛岀浜屽垪灏辨槸鍒氭墠鐨?ldquo;predictedpep”锛岀鍥涘垪鍒欐槸“bank-new.arff”涓師鏉ョ殑“pep”鍊硷紙杩欓噷閮芥槸“?”缂哄け鍊硷級銆傜涓夊垪瀵归娴嬬粨鏋滅殑缃俊搴︼紙confidence 锛夈€傛瘮濡傝瀵逛簬瀹炰緥0锛屾垜浠湁75%鐨勬妸鎻¤瀹冪殑“pep”鐨勫€间細鏄?ldquo;YES”锛屽瀹炰緥4鎴戜滑鏈?4.2%鐨勬妸鎻¤瀹冪殑“pep”鍊间細鏄?ldquo;NO”銆?br />鎴戜滑鐪嬪埌锛屼娇鐢ㄥ懡浠よ鑷冲皯鏈変袱涓ソ澶勩€備竴涓槸鍙互鎶婃ā鍨嬩繚瀛樹笅鏉ワ紝杩欐牱鏈夋柊鐨勫緟棰勬祴鏁版嵁鍑虹幇鏃讹紝涓嶇敤姣忔閲嶆柊寤烘ā锛岀洿鎺ュ簲鐢ㄤ繚瀛樺ソ鐨勬ā鍨嬪嵆鍙€傚彟涓€涓槸瀵归娴嬬粨鏋滅粰鍑轰簡缃俊搴︼紝鎴戜滑鍙互鏈夐€夋嫨鐨勯噰绾抽娴嬬粨鏋滐紝渚嬪锛屽彧鑰冭檻閭ｄ簺缃俊搴﹀湪85%浠ヤ笂鐨勭粨鏋溿€?br />
----鏁寸悊鑷?a href="http://maya.cs.depaul.edu/~classes/ect584/WEKA/classify.html" target="_blank">http://maya.cs.depaul.edu/~classes/ect584/WEKA/classify.html
6. 鑱氱被鍒嗘瀽
鍘熺悊涓庡疄鐜?nbsp;
鑱氱被鍒嗘瀽涓殑“绫?rdquo;锛坈luster锛夊拰鍓嶉潰鍒嗙被鐨?ldquo;绫?rdquo;锛坈lass锛夋槸涓嶅悓鐨勶紝瀵筩luster鏇村姞鍑嗙‘鐨勭炕璇戝簲璇ユ槸“绨?rdquo;銆傝仛绫荤殑浠诲姟鏄妸鎵€鏈夌殑瀹炰緥鍒嗛厤鍒拌嫢骞茬殑绨囷紝浣垮緱鍚屼竴涓皣鐨勫疄渚嬭仛闆嗗湪涓€涓皣涓績鐨勫懆鍥达紝瀹冧滑涔嬮棿璺濈鐨勬瘮杈冭繎锛涜€屼笉鍚岀皣瀹炰緥涔嬮棿鐨勮窛绂绘瘮杈冭繙銆傚浜庣敱鏁板€煎瀷灞炴€у埢鐢荤殑瀹炰緥鏉ヨ锛岃繖涓窛绂婚€氬父鎸囨姘忚窛绂汇€?br />鐜板湪鎴戜滑瀵瑰墠闈㈢殑“bank data”浣滆仛绫诲垎鏋愶紝浣跨敤鏈€甯歌鐨凨鍧囧€硷紙K-means锛夌畻娉曘€備笅闈㈡垜浠畝鍗曟弿杩颁竴涓婯鍧囧€艰仛绫荤殑姝ラ銆?br />K鍧囧€肩畻娉曢鍏堥殢鏈虹殑鎸囧畾K涓皣涓績銆傜劧鍚庯細1)灏嗘瘡涓疄渚嬪垎閰嶅埌璺濆畠鏈€杩戠殑绨囦腑蹇冿紝寰楀埌K涓皣锛?)璁″垎鍒绠楀悇绨囦腑鎵€鏈夊疄渚嬬殑鍧囧€硷紝鎶婂畠浠綔涓哄悇绨囨柊鐨勭皣涓績銆傞噸澶?)鍜?)锛岀洿鍒癒涓皣涓績鐨勪綅缃兘鍥哄畾锛岀皣鐨勫垎閰嶄篃鍥哄畾銆?br />
涓婅堪K鍧囧€肩畻娉曞彧鑳藉鐞嗘暟鍊煎瀷鐨勫睘鎬э紝閬囧埌鍒嗙被鍨嬬殑灞炴€ф椂瑕佹妸瀹冨彉涓鸿嫢骞蹭釜鍙栧€?鍜?鐨勫睘鎬с€俉EKA灏嗚嚜鍔ㄥ疄鏂借繖涓垎绫诲瀷鍒版暟鍊煎瀷鐨勫彉鎹紝鑰屼笖WEKA浼氳嚜鍔ㄥ鏁板€煎瀷鐨勬暟鎹綔鏍囧噯鍖栥€傚洜姝わ紝瀵逛簬鍘熷鏁版嵁“bank-data.csv”锛屾垜浠墍鍋氱殑棰勫鐞嗗彧鏄垹鍘诲睘鎬?ldquo;id”锛屼繚瀛樹负ARFF鏍煎紡鍚庯紝淇敼灞炴€?ldquo;children”涓哄垎绫诲瀷銆傝繖鏍峰緱鍒扮殑鏁版嵁鏂囦欢涓?ldquo;bank.arff”锛屽惈600鏉″疄渚嬨€?br />
鐢?ldquo;Explorer”鎵撳紑鍒氭墠寰楀埌鐨?ldquo;bank.arff”锛屽苟鍒囨崲鍒?ldquo;Cluster”銆傜偣“Choose”鎸夐挳閫夋嫨“SimpleKMeans”锛岃繖鏄疻EKA涓疄鐜癒鍧囧€肩殑绠楁硶銆傜偣鍑绘梺杈圭殑鏂囨湰妗嗭紝淇敼“numClusters”涓?锛岃鏄庢垜浠笇鏈涙妸杩?00鏉″疄渚嬭仛鎴?绫伙紝鍗矺=6銆備笅闈㈢殑“seed”鍙傛暟鏄璁剧疆涓€涓殢鏈虹瀛愶紝渚濇浜х敓涓€涓殢鏈烘暟锛岀敤鏉ュ緱鍒癒鍧囧€肩畻娉曚腑绗竴娆＄粰鍑虹殑K涓皣涓績鐨勪綅缃€傛垜浠笉濡ㄦ殏鏃惰瀹冨氨涓?0銆?br />閫変腑“Cluster Mode”鐨?ldquo;Use training set”锛岀偣鍑?ldquo;Start”鎸夐挳锛岃瀵熷彸杈?ldquo;Clusterer output”缁欏嚭鐨勮仛绫荤粨鏋溿€備篃鍙互鍦ㄥ乏涓嬭“Result list”涓繖娆′骇鐢熺殑缁撴灉涓婄偣鍙抽敭锛?ldquo;View in separate window”鍦ㄦ柊绐楀彛涓祻瑙堢粨鏋溿€?br />
缁撴灉瑙ｉ噴
棣栧厛鎴戜滑娉ㄦ剰鍒扮粨鏋滀腑鏈夎繖涔堜竴琛岋細
Within cluster sum of squared errors: 1604.7416693522332
杩欐槸璇勪环鑱氱被濂藉潖鐨勬爣鍑嗭紝鏁板€艰秺灏忚鏄庡悓涓€绨囧疄渚嬩箣闂寸殑璺濈瓒婂皬銆備篃璁镐綘寰楀埌鐨勬暟鍊间細涓嶄竴鏍凤紱瀹為檯涓婂鏋滄妸“seed”鍙傛暟鏀逛竴涓嬶紝寰楀埌鐨勮繖涓暟鍊煎氨鍙兘浼氫笉涓€鏍枫€傛垜浠簲璇ュ灏濊瘯鍑犱釜seed锛屽苟閲囩撼杩欎釜鏁板€兼渶灏忕殑閭ｄ釜缁撴灉銆備緥濡傛垜璁?ldquo;seed”鍙?00锛屽氨寰楀埌
Within cluster sum of squared errors: 1555.6241507629218
鎴戣鍙栧悗闈㈣繖涓€傚綋鐒跺啀灏濊瘯鍑犱釜seed锛岃繖涓暟鍊煎彲鑳戒細鏇村皬銆?nbsp;
鎺ヤ笅鏉?ldquo;Cluster centroids:”涔嬪悗鍒楀嚭浜嗗悇涓皣涓績鐨勪綅缃€?strong>瀵逛簬鏁板€煎瀷鐨勫睘鎬э紝绨囦腑蹇冨氨鏄畠鐨勫潎鍊硷紙Mean锛夛紱鍒嗙被鍨嬬殑灏辨槸瀹冪殑浼楁暟锛圡ode锛夛紝涔熷氨鏄杩欎釜灞炴€т笂鍙栧€间负浼楁暟鍊肩殑瀹炰緥鏈€澶氥€傚浜庢暟鍊煎瀷鐨勫睘鎬э紝杩樼粰鍑轰簡瀹冨湪鍚勪釜绨囬噷鐨勬爣鍑嗗樊锛圫td Devs锛夈€?nbsp;
鏈€鍚庣殑“Clustered Instances”鏄悇涓皣涓疄渚嬬殑鏁扮洰鍙婄櫨鍒嗘瘮銆?nbsp;
涓轰簡瑙傚療鍙鍖栫殑鑱氱被缁撴灉锛屾垜浠湪宸︿笅鏂?ldquo;Result list”鍒楀嚭鐨勭粨鏋滀笂鍙冲嚮锛岀偣“Visualize cluster assignments”銆傚脊鍑虹殑绐楀彛缁欏嚭浜嗗悇瀹炰緥鐨勬暎鐐瑰浘銆傛渶涓婃柟鐨勪袱涓鏄€夋嫨妯潗鏍囧拰绾靛潗鏍囷紝绗簩琛岀殑“color”鏄暎鐐瑰浘鐫€鑹茬殑渚濇嵁锛岄粯璁ゆ槸鏍规嵁涓嶅悓鐨勭皣“Cluster”缁欏疄渚嬫爣涓婁笉鍚岀殑棰滆壊銆?br />鍙互鍦ㄨ繖閲岀偣“Save”鎶婅仛绫荤粨鏋滀繚瀛樻垚ARFF鏂囦欢銆傚湪杩欎釜鏂扮殑ARFF鏂囦欢涓紝“instance_number”灞炴€ц〃绀烘煇瀹炰緥鐨勭紪鍙凤紝“Cluster”灞炴€ц〃绀鸿仛绫荤畻娉曠粰鍑虹殑璇ュ疄渚嬫墍鍦ㄧ殑绨囥€?br />
----鏁寸悊鑷?nbsp;http://maya.cs.depaul.edu/~classes/ect584/WEKA/k-means.html