"Loops / Iterations"

Armen

I consider the loops/iterations one of the most interesting tool of RapidMiner.

I am trying to use ClusterLoop, but something strange happened while using it.

In this example (see code) a ClusterLoop is set on the cluster-attribute: districtName. The ClusterLoop node contains a simple Replace node, which should “creates new attributes from nominal attributes with replaced substrings”: the Replace node is set on the attribute: dataset [“train”, “test”].

After setting the Debug-mode on, I start the process and check the results:

- the loop cycle accordingly to the different clusters, which is great!
- the values “test” are transformed into “TTTesTTT”, as expected
- the values “train” are transformed into “TTTesTTT” too…bug?

<操作符= " true " class = " select_attribute激活s" expanded="true" height="76" name="Select Attributes" width="90" x="179" y="165">
<参数键= "tribute_filter_type" value="subset"/>
<参数键= "tributes" value="Year|House_size_tot|House_size|Basement_size|House_ground|DistrictName|Price_2009|ID|dataset"/>








<参数键= "tribute_filter_type" value="single"/>
<参数键= "tribute" value="dataset"/>

In another process I work on each cluster applying a LinaerRegression. I want to see the results of the operation, but I am not able to show them in the “Result page”.
For that reason, I decided to save them using the nodes WritePerformance and WriteModel. In the path I need to use the Parameter Macros. It seems that they do not work (is ralated to this bug?http://bugs.rapid-i.com/show_bug.cgi?id=84#c0
I tried both the old version %{a} and the one proposed in the tutorial %{loop_value} , but they didn’t work!

<操作符= " true " class = " select_attribute激活s" expanded="true" height="76" name="Select Attributes" width="90" x="179" y="165">
<参数键= "tribute_filter_type" value="subset"/>
<参数键= "tributes" value="Year|House_size_tot|House_size|Basement_size|House_ground|DistrictName|Price_2009|ID|dataset"/>











<操作符= " true " class = " select_attribute激活s" expanded="true" height="76" name="Select Attributes (2)" width="90" x="313" y="30">
<参数键= "tribute_filter_type" value="subset"/>
<参数键= "tributes" value="Year|House_size_tot|House_size|Basement_size|House_ground|DistrictName|Price_2009|ID"/>






<操作符= " true " class = " select_attribute激活s" expanded="true" height="76" name="Select Attributes (3)" width="90" x="313" y="120">
<参数键= "tribute_filter_type" value="subset"/>
<参数键= "tributes" value="Year|House_size_tot|House_size|Basement_size|House_ground|DistrictName|Price_2009|ID"/>




<操作符= " true " class = " performance_reg激活ression" expanded="true" height="76" name="Performance" width="90" x="782" y="30">

land

Hi,
could you try to reproduce this behavior on a generated example set? Then I simply could load your process and see what's going wrong.
Thank you in advance.

Greetings,
Sebastian

Armen

Here is the data you can use to run these example streams:

ID;dataset;price;postcode;year;house_size;ground_size
1;train;4438979;A;1946;179;2759
2;train;9375083;B;1936;157;1986
3;test;8111665;C;1928;179;2428
4;train;8927810;D;1912;102;2872
5;test;643078;A;1996;102;1179
6;test;9116892;B;1920;177;1087
7;test;3343224;C;1918;177;1474
8;train;1559067;D;1999;141;2835
9;test;1642734;A;1991;156;1331
10;test;1143448;B;1998;103;1601
11;train;6054905;C;1913;95;1324
12;train;4652272;D;1938;141;1771
13;test;1007964;A;1987;188;1890
14;train;9427575;B;1999;115;964
15;train;3971974;C;1934;136;2155
16;test;9756122;D;1984;195;1759
17;test;6982381;A;1979;194;2855
18;train;5361267;B;1943;100;1740
19;test;92495;C;1997;93;2580
20;test;3853701;D;1997;126;2937
21;test;3672939;A;1934;105;2136
22;train;8553445;B;1984;110;939
23;train;2250150;C;1929;173;2851
24;test;2456895;D;1997;101;2952
25;test;5726626;A;1945;138;2499
26;test;9081445;B;1923;187;636
27个;测试;2280029;C; 1900; 187; 2812
28;test;2979554;D;1999;141;2743
29;test;3102467;A;1978;182;1706
30;test;6937219;B;1970;91;2273
31;train;5536152;C;1906;193;2424
32;train;8850263;D;1927;123;938
33;train;9247487;A;1920;154;2514
34;train;1245596;B;1931;90;2081
35;test;624383;C;1934;142;503
36;test;5707598;D;1977;192;2749
37;train;4695273;A;1976;155;1156
38;train;5992180;B;1958;108;2907
39;test;4963939;C;1911;198;1493
40;test;8146548;D;1928;125;2466
41;test;927543;A;1900;196;1125
42;test;9800296;B;1925;143;2194
43;train;5982102;C;1955;158;565
44;train;8238828;D;1936;196;1127
45;train;7748803;A;1980;191;539
46;train;1028282;B;1964;200;705
47;test;6424228;C;1912;162;2277
48;train;9347885;D;1980;195;1902
49;train;5836803;A;1976;184;2611
50;train;6594397;B;1909;108;1599

Here is the example code for the Replace problem:

<参数键= "tribute_filter_type" value="single"/>
<参数键= "tribute" value="dataset"/>

Here is the example code for the save problem:

<操作符= " true " class = " select_attribute激活s" expanded="true" height="76" name="Select Attributes (2)" width="90" x="313" y="30">
<参数键= "tribute_filter_type" value="subset"/>
<参数键= "tributes" value="ID|year|price|postcode|house_size|ground_size"/>






<操作符= " true " class = " select_attribute激活s" expanded="true" height="76" name="Select Attributes (3)" width="90" x="313" y="120">
<参数键= "tribute_filter_type" value="subset"/>
<参数键= "tributes" value="ID|year|price|postcode|house_size|ground_size"/>




<操作符= " true " class = " performance_reg激活ression" expanded="true" height="76" name="Performance" width="90" x="782" y="30">

land

Hi,
everything works fine for me, so every bug seems to be already fixed in the current version.
So your problems will be solved with the final version.

Greetings,
Sebastian