以LibSvm爲例給出weka輸出結果的簡單翻譯 [WEKA使用]
post by 河內塔 / 2012-1-11 15:51 Wednesday
中午在百度上看到,一個網友提問了個問題,翻譯weka執行LibSvm的輸出結果,調用的數據來自weka自帶的數據包中的glass.arff文件,大家都知道咱們使用的weka都是英文版的,看看確實有些地方還真不知道怎麼翻譯,想想已經學了半年了,總該把輸出結果弄清楚吧,於是硬着頭皮,一點點搜索、查資料、搜索來翻譯,不過總算功夫不負有心人,花了一個多小時,終於翻譯完了。下面貼出來,大家共享!
具體的操作步驟就不多說了,如果還不會,可以參照我前面的兩篇博文weka如何使用LivSvm:http://www.heneita.com/?post=14和如何用k-means進行聚類:http://www.heneita.com/?post=10。
原文:
=== Run information ===
Relation: Glass
Instances:214
Attributes:10
RI
Na
Mg
Al
Si
K
Ca
Ba
Fe
Type
Test mode:10-fold cross-validation
=== Classifier model (full training set) ===
LibSVM wrapper, original code by Yasser EL-Manzalawy (= WLSVM)
Time taken to build model: 0.05seconds
=== Stratified cross-validation ===
=== Summary ===
Correctly Classified Instances 147 68.6916 %
Incorrectly Classified Instances 67 31.3084 %
Kappa statistic 0.5555
Mean absolute error 0.0895
Root mean squared error 0.2991
Relative absolute error 42.2426 %
Root relative squared error 92.1555 %
Total Number of Instances 214
=== Detailed Accuracy By Class ===
TP Rate FP Rate Precision Recall F-Measure ROC Area Class
0.814 0.215 0.648 0.814 0.722 0.8 build wind float
0.737 0.21 0.659 0.737 0.696 0.763 build wind non-float
0 0 0 0 0 0.5 vehic wind float
0 0 0 0 0 ? vehic wind non-float
0.692 0.015 0.75 0.692 0.72 0.839 containers
0.222 0.01 0.5 0.222 0.308 0.606 tableware
0.793 0.011 0.92 0.793 0.852 0.891 headlamps
Weighted Avg. 0.687 0.148 0.637 0.687 0.655 0.77
=== Confusion Matrix ===
a b c d e f g <-- classified as
57 13 0 0 0 0 0 | a = build wind float
17 56 0 0 2 1 0 | b = build wind non-float
12 5 0 0 0 0 0 | c = vehic wind float
0 0 0 0 0 0 0 | d = vehic wind non-float
0 3 0 0 9 0 1 | e = containers
1 4 0 0 1 2 1 | f = tableware
1 4 0 0 0 1 23 | g = headlamps
下面是翻譯的結果:
=== Run information ===運行信息
Scheme(模型):weka.classifiers.functions.LibSVM -S 0 -K 2 -D 3 -G 0.0 -R 0.0 -N 0.5 -M 40.0 -C 1.0 -E 0.0010 -P 0.1 注:這裏用的是libsvm模型,調用的是weka.classifiers.function.libsvm包
Relation(關係表,類似數據庫關係表,名字和arff文件名字一樣): Glass(玻璃)
Instances(實例數,每個玻璃的一組數據表示一個實例): 214
Attributes(屬性): 10
RI(玻璃折射率)
Na(鈉元素)
Mg(鎂)
Al (鋁)
Si (硅)
K (鉀)
Ca (鈣)
Ba (鋇)
Fe (鐵)
Type (類別,人工加上的玻璃的類標)
Test mode:10-fold cross-validation 測試模式:10摺疊交叉驗證(用來驗證模型結果的穩定性)
=== Classifier model (full training set) ===分類模型(整個數據集作爲訓練集)
LibSVM wrapper, original code by Yasser EL-Manzalawy (= WLSVM)(封裝的livsvm源碼來自一個叫EL-Manzalawy的人)
Time taken to build model: 0.02 seconds(建立模型(也叫分類器)所需的時間):0.02秒
=== Stratified cross-validation ===分層交叉驗證(這裏這種驗證方法沒有用)
=== Summary ===結果總結
Correctly Classified Instances(正確分類的實例) 148 69.1589 %
Incorrectly Classified Instances(錯誤分類的實例) 66 30.8411 %
Kappa statistic(Kappa統計量) 0.3579
Mean absolute error(均值絕對誤差) 0.0881
Root mean squared error(均方根誤差) 0.2968
Relative absolute error(相對絕對誤差) 60.7715 %
Root relative squared error(相對均方根誤差) 111.5949 %
Total Number of Instances(參與實驗的實例(數據)總數) 214
=== Detailed Accuracy By Class ===分類的具體精度
TP Rate(真正元比率) FP Rate(假正元比率) Precision(查準率) Recall(查全率) F-Measure(F-測量,直說吧查詢率與查全率的調和平均值) ROC Area(ROC曲線下方的面積,表示精度) Class(類別)
0.847 0.5 0.676 0.847 0.752 0.674 build wind float(玻璃種類)
0.5 0.153 0.727 0.5 0.593 0.674 build wind non-float(玻璃種類)
0 0 0 0 0 ? vehic wind float(玻璃種類)
0 0 0 0 0 ? vehic wind non-float(玻璃種類)
0 0 0 0 0 ? containers(玻璃種類)
0 0 0 0 0 ? tableware(玻璃種類)
0 0 0 0 0 ? headlamps(玻璃種類)
Weighted Avg(加權平均). 0.692 0.344 0.699 0.692 0.68 0.674
=== Confusion Matrix ===(混淆矩陣)
a b c d e f g <-- classified as(下面都是玻璃的類別,開參見glass.arff文件)
100 18 0 0 0 0 0 | a = build wind float(玻璃種類)
48 48 0 0 0 0 0 | b = build wind non-float(玻璃種類)
0 0 0 0 0 0 0 | c = vehic wind float(玻璃種類)
0 0 0 0 0 0 0 | d = vehic wind non-float(玻璃種類)
0 0 0 0 0 0 0 | e = containers(玻璃種類)
0 0 0 0 0 0 0 | f = tableware(玻璃種類)
0 0 0 0 0 0 0 | g = headlamps(玻璃種類)