K-NN - Platform For AI - Alibaba Cloud Documentation Center

The K-NN component selects the K-nearest records from a row in the prediction table for classification. The most common class of the K-nearest records is used as the class of the row.

Configure the component

You can use one of the following methods to configure the K-NN component.

Method 1: Configure the component on the pipeline page

You can configure the parameters of the K-NN component on the pipeline page of Machine Learning Designer of Machine Learning Platform for AI (PAI). Machine Learning Designer is formerly known as Machine Learning Studio. The following table describes the parameters.

Tab	Parameter	Description
Fields Setting	Feature Columns in the Training Table	The feature columns that are used for training.
	Specifies the label column in the training table	The label column that is selected for training.
	Feature Columns in the Prediction Table	If this parameter is not specified, the feature columns selected from the prediction table are the same as the feature columns in the training table.
	Append ID Column to Output Table	The ID columns that are used to obtain the predicted values of a column. By default, the feature columns selected from the prediction table are used as the ID columns.
	Input in Sparse Format	Specifies whether the input data is in the sparse format. If you select the check box, the input data is in the key-value format.
	KV Pair Delimiter	The delimiter that is used to separate key-value pairs. Commas (,) are used by default.
	Key and Value Delimiter	The delimiter that is used to separate keys and values. Colons (:) are used by default.
Parameters Setting	Number of Neighbors	Default value: 100.
Tuning	Number of Cores	The number of cores. By default, the system determines the value.
Tuning	Memory Size	The memory size of each core. By default, the system determines the value.

Method 2: Use PAI commands

Configure the component parameters by using PAI commands. You can use the SQL Script component to call PAI commands. For more information, see SQL Script.

PAI -name knn
    -DtrainTableName=pai_knn_test_input
    -DtrainFeatureColNames=f0,f1
    -DtrainLabelColName=class
    -DpredictTableName=pai_knn_test_input
    -DpredictFeatureColNames=f0,f1
    -DoutputTableName=pai_knn_test_output
    -Dk=2;

Parameter	Required	Description	Default value
trainTableName	Yes	The name of the training table.	N/A
trainFeatureColNames	Yes	The names of the feature columns in the training table.	N/A
trainLabelColName	Yes	The name of the label column in the training table.	N/A
trainTablePartitions	No	The partitions that are selected from the training table for training.	All partitions
predictTableName	Yes	The name of the prediction table.	N/A
outputTableName	Yes	The name of the output table.	N/A
predictFeatureColNames	No	The names of the feature columns in the prediction table.	Same as the value of the trainFeatureColNames parameter
predictTablePartitions	No	The partitions that are selected from the prediction table for prediction.	All partitions
appendColNames	No	The names of the columns appended to the output table.	Same as the value of the predictFeatureColNames parameter
outputTablePartition	No	The partitions in the output table.	Full table
k	No	The number of K-nearest neighbors. Valid values: 1 to 1000.	100
enableSparse	No	Specifies whether data in the input table is in the sparse format. Valid values: true and false.	false
itemDelimiter	No	The delimiter that is used to separate key-value pairs.	,
kvDelimiter	No	The delimiter that is used to separate the key and value in a key-value pair.	:
coreNum	No	The number of cores. This parameter must be used with the memSizePerCore parameter. Valid values: 1 to 20000.	Determined by the system
memSizePerCore	No	The memory size of each core. Valid values: 1024 to 64 x 1024. Unit: MB.	Determined by the system
lifecycle	No	The lifecycle of the output table.	N/A

Example

Generate the training data.

create table pai_knn_test_input as
select * from
(
  select 1 as f0,2 as f1, 'good' as class
  union all
  select 1 as f0,3 as f1, 'good' as class
  union all
  select 1 as f0,4 as f1, 'bad' as class
  union all
  select 0 as f0,3 as f1, 'good' as class
  union all
  select 0 as f0,4 as f1, 'bad' as class
)tmp;

Run the following PAI command to submit the parameters of the K-NN component:

pai -name knn
    -DtrainTableName=pai_knn_test_input
    -DtrainFeatureColNames=f0,f1
    -DtrainLabelColName=class
    -DpredictTableName=pai_knn_test_input
    -DpredictFeatureColNames=f0,f1
    -DoutputTableName=pai_knn_test_output
    -Dk=2;

View the training result. The result contains the following columns:
- f0 and f1: the appended columns.
- prediction_result: lists the classification results.
- prediction_score: lists the probabilities for the classification results.
- prediction_detail: lists the classes of the K-nearest neighbors and their probabilities.