數(shù)據(jù)挖掘SPSSclementine培訓(xùn)PPT學(xué)習(xí)教案

1、會計學(xué)1數(shù)據(jù)挖掘數(shù)據(jù)挖掘SPSSclementine培訓(xùn)培訓(xùn)2第1頁/共83頁第2頁/共83頁4第3頁/共83頁5第4頁/共83頁6第5頁/共83頁7第6頁/共83頁8第7頁/共83頁9精煉模型未精煉模型第8頁/共83頁10決定一個數(shù)值或分類結(jié)果第9頁/共83頁11輸出隱藏層輸入層第10頁/共83頁12Cat.%nBad52.01168Good47.99155Total (100.00) 323Credit ranking (1=default)Cat.%nBad86.67143Good13.3322Total (51.08) 165Paid Weekly/MonthlyP-value=0.0

2、000, Chi-square=179.6665, df=1Weekly payCat.%nBad15.8225Good84.18133Total (48.92) 158Monthly salaryCat.%nBad90.51143Good9.4915Total (48.92) 158Age CategoricalP-value=0.0000, Chi-square=30.1113, df=1Young ( 35)Cat.%nBad48.9824Good51.0225Total (15.17) 49Age CategoricalP-value=0.0000, Chi-square=58.725

3、5, df=1Young ( 35)Cat.%nBad0.000Good 100.008Total(2.48)8Social ClassP-value=0.0016, Chi-square=12.0388, df=1Management;ClericalCat.%nBad58.5424Good41.4617Total (12.69) 41Professional第11頁/共83頁13第12頁/共83頁14第13頁/共83頁15第14頁/共83頁16第15頁/共83頁17第16頁/共83頁18第17頁/共83頁19第18頁/共83頁20和和第19頁/共83頁21第20頁/共83頁22第21頁/共

4、83頁23第22頁/共83頁24第23頁/共83頁25第24頁/共83頁26第25頁/共83頁27第26頁/共83頁28第27頁/共83頁29第28頁/共83頁30第29頁/共83頁31第30頁/共83頁32第31頁/共83頁33第32頁/共83頁34第33頁/共83頁35第34頁/共83頁36第35頁/共83頁37第36頁/共83頁38第37頁/共83頁39第38頁/共83頁40第39頁/共83頁41第40頁/共83頁42n保存更新的流第41頁/共83頁43第42頁/共83頁44第43頁/共83頁45第44頁/共83頁46第45頁/共83頁47第46頁/共83頁48第47頁/共83頁49第4

5、8頁/共83頁50第49頁/共83頁51第50頁/共83頁52第51頁/共83頁53第52頁/共83頁54第53頁/共83頁55第54頁/共83頁56第55頁/共83頁57第56頁/共83頁58第57頁/共83頁59第58頁/共83頁60第59頁/共83頁61第60頁/共83頁62第61頁/共83頁63第62頁/共83頁64第63頁/共83頁65第64頁/共83頁66第65頁/共83頁67第66頁/共83頁68第67頁/共83頁69第68頁/共83頁70第69頁/共83頁71n關(guān)聯(lián)規(guī)則實例僅僅是相配前提的記錄n序列規(guī)則實例相配整個規(guī)則的記錄第70頁/共83頁72第71頁/共83頁73第72頁/

6、共83頁74第73頁/共83頁75第74頁/共83頁76第75頁/共83頁77 1. Perform many operations in the database.2. Perform the rest on a powerfulserver.3. Use the client processor for viewing results.4. Data isnt passedacross the networkunnecessarily.Data PreparationClementine FeaturesFull-process in-database mining with SQL op

7、timizationScalable performanceData PreparationClementine Features第76頁/共83頁78第77頁/共83頁79 CEMIClementine FeaturesNeed something else?Add new algorithms, data preparation, graphics, maps call other applications, etc even those that have been developed in-house第78頁/共83頁80 Clementine Stream ScriptNotes:F

8、or loop using in_fields.Loop applies to IN fields only.Train net node changes name so refer to it by type (theres only one).As usual, script “goes with” stream.# Start with empty palette.clear generated palette# Now loop through all IN fields and# train with each in turn set to NONE.# Give models appropriate names.for f in_fields_at typeif type.direction.f = INset type.direction.f = NONEset :name = omit_ fexecute :trainnetset