how to get the ranks of unlabeled case using K-NN

inceptorfullinceptorfull Member Posts: 44 Contributor I
edited November 2018 in Help
Hi all, I have unlabeled data and want to get rank of its nearest cases so I can compare it with them, Its credit rating problems so I have unlabeled customers and want to know the nearest neighbor of them by ranking or how close they are to the good or bad customers

Answers

  • JEdwardJEdward RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 578 Unicorn
    It sounds like you aren't needing the k-NN operator, but rather the Cross Distances.  (Other similarity operators are also useable). 
    Have a look at this sample process using the Golf dataset. 
    <?xml version="1.0" encoding="UTF-8" standalone="no"?>
    <process version="7.0.000">
      <context>
        <input/>
        <output/>
        <macros/>
      </context>
      <operator activated="true" class="process" compatibility="6.0.002" expanded="true" name="Process">
        <process expanded="true">
          <operator activated="true" class="retrieve" compatibility="7.0.000" expanded="true" height="68" name="Golf" width="90" x="45" y="187">
            <parameter key="repository_entry" value="//Samples/data/Golf"/>
          </operator>
          <operator activated="true" class="generate_id" compatibility="7.0.000" expanded="true" height="82" name="Generate ID" width="90" x="179" y="238">
            <parameter key="create_nominal_ids" value="true"/>
            <description align="center" color="transparent" colored="false" width="126">Using nominal IDs just to demo.</description>
          </operator>
          <operator activated="true" class="retrieve" compatibility="7.0.000" expanded="true" height="68" name="Retrieve Golf-Testset" width="90" x="45" y="34">
            <parameter key="repository_entry" value="//Samples/data/Golf-Testset"/>
          </operator>
          <operator activated="true" class="subprocess" compatibility="7.0.000" expanded="true" height="82" name="Get only 1 record." width="90" x="179" y="34">
            <process expanded="true">
              <operator activated="true" class="sample" compatibility="7.0.000" expanded="true" height="82" name="Sample" width="90" x="45" y="34">
                <parameter key="sample_size" value="1"/>
                <list key="sample_size_per_class"/>
                <list key="sample_ratio_per_class"/>
                <list key="sample_probability_per_class"/>
              </operator>
              <operator activated="true" class="select_attributes" compatibility="7.0.000" expanded="true" height="82" name="Select Attributes" width="90" x="179" y="34">
                <parameter key="attribute_filter_type" value="single"/>
                <parameter key="attribute" value="Play"/>
                <parameter key="invert_selection" value="true"/>
                <parameter key="include_special_attributes" value="true"/>
              </operator>
              <operator activated="true" class="generate_id" compatibility="7.0.000" expanded="true" height="82" name="Generate ID (2)" width="90" x="313" y="34"/>
              <connect from_port="in 1" to_op="Sample" to_port="example set input"/>
              <connect from_op="Sample" from_port="example set output" to_op="Select Attributes" to_port="example set input"/>
              <connect from_op="Select Attributes" from_port="example set output" to_op="Generate ID (2)" to_port="example set input"/>
              <connect from_op="Generate ID (2)" from_port="example set output" to_port="out 1"/>
              <portSpacing port="source_in 1" spacing="0"/>
              <portSpacing port="source_in 2" spacing="0"/>
              <portSpacing port="sink_out 1" spacing="0"/>
              <portSpacing port="sink_out 2" spacing="0"/>
            </process>
          </operator>
          <operator activated="true" class="cross_distances" compatibility="7.0.000" expanded="true" height="103" name="Cross Distances" width="90" x="313" y="85">
            <parameter key="only_top_k" value="true"/>
            <parameter key="k" value="3"/>
          </operator>
          <operator activated="true" class="select_attributes" compatibility="7.0.000" expanded="true" height="82" name="Select only label" width="90" x="447" y="238">
            <parameter key="attribute_filter_type" value="subset"/>
            <parameter key="attributes" value="id|Play"/>
            <parameter key="include_special_attributes" value="true"/>
          </operator>
          <operator activated="true" class="join" compatibility="7.0.000" expanded="true" height="82" name="Join to Request" width="90" x="447" y="34">
            <parameter key="use_id_attribute_as_key" value="false"/>
            <list key="key_attributes">
              <parameter key="request" value="id"/>
            </list>
            <description align="center" color="transparent" colored="false" width="126">This join is just to get the original data back rather than just the ID.</description>
          </operator>
          <operator activated="true" class="join" compatibility="7.0.000" expanded="true" height="82" name="Join to Reference" width="90" x="581" y="187">
            <parameter key="use_id_attribute_as_key" value="false"/>
            <list key="key_attributes">
              <parameter key="document" value="id"/>
            </list>
            <description align="center" color="transparent" colored="false" width="126">When the result is joined with the original Reference dataset then the label is used.</description>
          </operator>
          <connect from_op="Golf" from_port="output" to_op="Generate ID" to_port="example set input"/>
          <connect from_op="Generate ID" from_port="example set output" to_op="Cross Distances" to_port="reference set"/>
          <connect from_op="Retrieve Golf-Testset" from_port="output" to_op="Get only 1 record." to_port="in 1"/>
          <connect from_op="Get only 1 record." from_port="out 1" to_op="Cross Distances" to_port="request set"/>
          <connect from_op="Cross Distances" from_port="result set" to_op="Join to Request" to_port="left"/>
          <connect from_op="Cross Distances" from_port="request set" to_op="Join to Request" to_port="right"/>
          <connect from_op="Cross Distances" from_port="reference set" to_op="Select only label" to_port="example set input"/>
          <connect from_op="Select only label" from_port="example set output" to_op="Join to Reference" to_port="right"/>
          <connect from_op="Join to Request" from_port="join" to_op="Join to Reference" to_port="left"/>
          <connect from_op="Join to Reference" from_port="join" to_port="result 1"/>
          <portSpacing port="source_input 1" spacing="0"/>
          <portSpacing port="sink_result 1" spacing="0"/>
          <portSpacing port="sink_result 2" spacing="0"/>
        </process>
      </operator>
    </process>
  • inceptorfullinceptorfull Member Posts: 44 Contributor I
    thanks a lot for quick reply and help, I will try to get more understanding of it and will give you feedback, but want to know is that based on the K-NN? as i can see the number of neighours and distance so on

    also if you have more tutorial on that process i will be pleased to tell me, I will give you feedback soon thanks again apperciate it
  • inceptorfullinceptorfull Member Posts: 44 Contributor I
    it just give me the distance , I donot distance from what? also I have 516 case so I found huge distance

    I want to enter the unlabeled cases to be assigned for the most close similar case, using nearest neighbour, I donot know how to do it
    it is something like that

    https://dato.com/learn/userguide/nearest_neighbors/nearest_neighbors.html

    thanks a gain
  • JEdwardJEdward RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 578 Unicorn
    If you are wanting to assign it to the value of the nearest value then k-NN with k = 1 is what you are looking for.  If it is to look at the nearest 3 cases then k-NN with k = 3 is what you would like, this would assign the missing labels to the closest record where you do have a value by weight vote. 
    As you are assigning it to the missing labels, maybe try the operator 'Impute Missing Values' with k-NN inside it. 

    If what you are looking is what the k closest records to your sample record is then the similarity operators (such as Cross Distance) are what you need. 

    What do you want to happen in your process? 
    :)
  • inceptorfullinceptorfull Member Posts: 44 Contributor I
    am really thankfull for your feedback and keeping up with me, actually my last step in my research depends on that step so hope to help me,

    first of all, I want to enter training data to make the model train on ( Neural network or K-nn) whatever is okie,

    then enter the unlableled data ( same as exampleset but with missing values in the label column)

    the result to give me the best 5 closest and similar cases from the labeled data ( Exampleset) , so I donot know the right operator to use, secondly the results appear like that using the cross distance

    image

    but i want it to appear in something like that ( i used spss modeler but there isno predication in it )
    image
Sign In or Register to comment.