Exploration API

About the Exploration API

The Exploration API allows for the request of statistical metrics of SymetryML projects using univariate, bivariate, Hypothesis Testing (ztest, test, ftest, ANOVA) as well as Information Gain analyses. The body of the request contains a ExploreContext which in turn is a list of MLContext, one for each feature for which exploration data is needed.

MLContext Parameters for Peer Exploration

If the project is part of a federation it might be possible to ask that the exploration result is fetched from a given node. This can be done by specifying which peer to use with the following extra parameters inside a MLContext. Please consult the About Federated Learning : Peer Exploration section for details.

Key
fed_peer_uuid_for_explore	Peer ID to use for exploration results.

URL

POST /symetry/rest/{cid}/projects/{pid}/explore?metric={_metric_} [body=ExploreContext]

Query Parameters

Parameter	Required / Optional	Description
metric	Required	Metric to use for exploration. For more information, see the section List of Metrics
async	Optional	If set to true then the exploration will be done asynchronously and the result will be fetched using the Job API.
returnProjectInfo	Optional	Whether to return project info in the response. Default is `true`.

HTTP Responses

HTTP Status Code	HTTP Status Message	Description
200	OK	Success.
400	BAD REQUEST	Unknown SymetryML project. `{"statusCode":"BAD_REQUEST","statusString":" Cannot Find SYMETRYML id[r2] for Customer id [c1]","values":{}}`

HTTP Response Entity

The response is a KSVDMap entity. This structure is an array of map<key,value>, where keys are string and values are an Institute of Electrical and Electronics Engineers (IEEE) 64-bit double floating-point number. One such map is added to the response for each MLContext in the ExploreContext request body. The keys in each map depend on the metric query parameter. For more information, see the section ExploreContext.

HTTP Response Entity	Example
KSVDMap	`{"statusCode":"OK","statusString":"OK","values":{"KSVDMap":{"values":[{"count":150.0,"mean":5.843333333333334,"variance":0.6811222222222204,"skewness":0.31808651443435426,"stddev":0.8253012917851398}]}}}`
ProjectInfo	A project info entity is also returned to reflect the state of the project that might have changed.

List of "Metric" Query Parameters

Depending on the metric query parameter, different information is returned in the KSVDMap JSON response entity. The KSVDMap JSON entity is an array of map<key, value>, where key is a string and value is an IEEE 64-bit double floating-point number. The following table enumerates the keys that the response will contain for all valid metric query parameters.

Metric	Description	KSVDMap Key
uni	Returns all univariate metrics for the attribute listed in each MLContext in the ExploreContext. If an MLContext has more than one attribute, only the first one is used.	- count = count of the attribute. - mean = mean of the attribute. - variance = variance of the attribute. - stddev = standard deviation of the attribute. - skewness = skewness of the attribute. - kurtosis = kurtosis of the attribute. - stderr = the standard deviation divided by the square root of the count for this feature. - meanMarginErrorC95 = Mean standard error 95% confidence level.
uni (with histogram)	For project with histogram enabled, additional information is returned. Please consult this section to learn how to enable histogram for a project.	- min = minimum value for this feature. - max = maximum value for this feature. - median = median value for this feature.
bi	Returns all bivariate metrics for the two attributes listed in each MLContext in the ExploreContext. Each MLContext should contain at least two attribute indexes. If there are more than two, only the first two are used.	- covar = covariance between the two attributes. - linCorr = linear correlation between the two attributes. - condMean = conditional mean. - condCount = conditional count. - condVariance = conditional variance. - condSum = conditional sum. - condStddev = conditional standard deviation. - biCi95Down = bivariate 95% confidence interval lower bound. - biCi95Up = bivariate 95% confidence interval upper bound.
ztest	Returns the z-test metrics using all of the MLContext in the ExploreContext request body.	- zn1 = count between attribute and target 1. - zn2 = count between attribute and target 2. - zm1 = mean between attribute and target 1. - zm2 = mean between attribute and target 2. - zs1 = variance between attribute and target 1. - zs2 = variance between attribute and target 2. - z = z test value. - zp = z test value probability.
ztestp	Returns the z-test-proportion metrics using all of the MLContext in the ExploreContext request body.	- z = z test value. - zp = z test value probability.
ztestmu	Returns the z-test - against a known mean - metrics using all of the MLContext in the ExploreContext request body.	- z = z test value. - zp = z test value probability.
ftest	Returns the f-test metrics using all of the MLContext in the ExploreContext request body.	- fn1 = count between attribute and target 1. - fn2 = count between attribute and target 2. - fm1 = mean between attribute and target 1. - fm2 = mean between attribute and target 2. - fs1 = variance between attribute and target 1. - fs2 = variance between attribute and target 2. - fdf1 = degree of freedom between attribute and target 1. - fdf2 = degree of freedom between attribute and target 2. - f = F test value. - fp = F test value probability.
ftestsigma	Returns the f-test - against a known variance - metrics using all of the MLContext in the ExploreContext request body.	- f = F test value. - fp = F test value probability.
ttest	Returns the t-test metrics using all of the MLContext in the ExploreContext request body.	- tn1 = count between attribute and target 1. - tn2 = count between attribute and target 2. - tm1 = mean between attribute and target 1. - tm2 = mean between attribute and target 2. - ts1 = variance between attribute and target 1. - ts2 = variance between attribute and target 2. - t = t test value. - tdf = t test value degree of freedom. - tp = t test value probability. - tu = t test value for unequal variance. - tdfu = t test value degree of freedom for unequal variance. - tpu = t test value probability for unequal variance.
ttestmu	Returns the T-test - against a known mean - metrics using all of the MLContext in the ExploreContext request body.	- t = T test value. - tp = T test value probability.
anova	Returns the ANOVA metrics using all of the MLContext in the ExploreContext request body.	- SSb = sum square between groups. - SSw = sum square within groups. - Dfb = degree of freedom between groups. - Dfw = degree of freedom within groups. - MSb = mean square between groups. - MSw = mean square within groups. - F = F value from F distribution. - Ssgamma = total sum of square. - Dfgamma* = total degree of freedom. - P = probability of F.
chi2	Returns chi-square information using all of the MLContext in the ExploreContext request body.	- chiStat = chi-square value. - df = degree of freedom. - pval = probability. - coef = contingency coefficient. Additionally, for all the pair wise values the observed count will be returned as a value using the following format for the key. Example: obs$1:4
gain	Returns information gain for a given attribute –input - given another attributes – target.

Note ztestmu, ttestmu, and ftestsigma require the following parameters to be set in MLContext.extraParameters respectively: sml_explore_ztest_known_mu, sml_explore_ttest_known_mu, and sml_explore_ftest_known_sigma.

Various Hypothesis Test

Its possible to perform variations on ztest, ftest, ttest, anova, and chi-squared test. Depending on which test needs to be performed the MLContext must be populated differently.

Hypothesis Testing Suite
Hypothesis Test	MLContext Content
Z test Compare means of 2 continuous features.	`inputAttributes`: contains 2 continuous attributes ids
Z test Compare mean of 1 continuous feature against a known mean	* `inputAttributes`: contains 1 continuous attribute id * `extraParameters`: contains one key `sml_explore_ztest_known_mu` with known mean
Z test Compare means of a continuous feature conditioned on 2 binary features	`inputAttributes`: contains 3 attribute id, first one is continuous followed by 2 binary ids
Z test proportion Compare two proportions originated from two binary features	`inputAttributes`: contains 2 binary attribute id
F test Compare variance of 2 continuous features.	`inputAttributes`: contains 2 continuous attributes id
F test Compare variance of continuous feature against a known variance	* `inputAttributes`: contains 1 continuous attribute id * `extraParameters`: contains one key `sml_explore_ftest_known_sigma` with known variance / sigma
F test Compare variances of a continuous feature conditioned on 2 binary features	`inputAttributes`: contains 3 attribute id, first one is continuous followed by 2 binary ids
T test Compare means of 2 continuous features.	`inputAttributes`: contains 2 continuous attributes id
T test Compare mean of 1 continuous feature against a known mean	* `inputAttributes`: contains 1 continuous attribute id * `extraParameters`: contains one key `sml_explore_ttest_known_mu` with known mean
T test Compare means of a continuous feature conditioned on 2 binary features	`inputAttributes`: contains 3 attribute id, first one is continuous followed by 2 binary ids
Anova Compare means of 2 or more continuous features	`inputAttributes`: contains 2 or more continuous attribute id
Anova Compare means of a continuous feature conditioned on 2 or more binary features	* `inputAttributes`: contains 1 continuous attribute id * `targets`: contains 1 or more binary attribute id
Chi square Determine the association between binary features.	`inputAttributes`: contains 2 or more binary attribute ids `targets`: contains 2 or more binary attribute ids see Chi Square for example.
Chi square Determine the association between binary features based on category names	`inputAttributes`: list of category name See Chi Square for example.

MLContext Extra Parameters for Bivariate Statistics

When using the bi metric, the following extra parameters can be included in the MLContext extraParameters field to request additional conditional statistics:

Parameter	Type	Description
need_bi_conditional	Boolean	Request bivariate conditional statistics
need_bi_conditional_mmm	Boolean	Request conditional min, max, and median values
need_bi_conditional_ci95	Boolean	Request conditional 95% confidence interval bounds

Sample Request/Response Bivariate Statistics

Request:
POST url="http://charm:8080/symetry/rest/c1/projects/irisTest/explore?metric=bi"

Body:
{"values":[{"targets":[],"inputAttributes":["0","3"],"extraParameters":{}}]}

Response:
{"statusCode":"OK","statusString":"OK","values":{"smlInfo":{"pid":"r1","isDirty":true,"modelsList":[],"modelTypeList":[],"attributeNames":["sepal_length","sepal_width","petal_length","petal_width","sepal_lengt_b1","sepal_lengt_b2","sepal_width_b1","sepal_width_b2","petal_length_b1","petal_length_b2","petal_width_b1","petal_width_b2","Iris_setosa","Iris_versicolor","Iris_virginica"],"attributeIndexes":[0,1,2,3,4,5,6,7,8,9,10,11,12,13,14],"attributeTypes":["C","C","C","C","B","B","B","B","B","B","B","B","B","B","B"],"modelAssessments":{},"modelPredictions":{},"hash":-1,"categorySeparator":"$","type":"cpu","creationDate":1488220762857,"lastModificationDate":1488220780484,"loaded":true,"persisted":true},"KSVDMap":{"values":[{"linCorr":0.8180000000000001,"condVariance":30.9294,"covar":0.5135000000000001,"condCount":179.8,"condMean":6.2717}]}}}

Sample Request/Response chi2 Example 1

Request:
POST url="http://charm:8080/symetry/rest/c1/projects/irisTest/explore?metric=chi2"

Body:
{"values":[{"targets":["12","13","14"],"inputAttributes":["10","11"],"inputAttributeNames":[],"extraParameters":{}}]}

Response:
{"statusCode":"OK","statusString":"OK","values":{"smlInfo":{"pid":"irisP","isDirty":true,"modelsList":[],"modelTypeList":[],"dsList":["Iris_rtlm.csv"],"attributeNames":["sepal_length","sepal_width","petal_length","petal_width","sepal_lengt_b1","sepal_lengt_b2","sepal_width_b1","sepal_width_b2","petal_length_b1","petal_length_b2","petal_width_b1","petal_width_b2","Iris_setosa","Iris_versicolor","Iris_virginica"],"attributeIndexes":[0,1,2,3,4,5,6,7,8,9,10,11,12,13,14],"attributeTypes":["C","C","C","C","B","B","B","B","B","B","B","B","B","B","B"],"modelAssessments":{},"modelPredictions":{},"params":{},"hash":-1,"categorySeparator":"$","type":"cpu","creationDate":1588623147023,"lastModificationDate":1588623148044,"loaded":true,"persisted":true,"histogramEnabled":true},"KSVDMap":{"values":[{"df":2.0,"obs$11:12":0.0,"coef":0.8409000000000001,"pval":0.0,"chiStat":150.0,"obs$10:14":0.0,"obs$10:13":0.0,"obs$10:12":50.0,"obs$11:13":50.0,"obs$11:14":50.0}]}}}

Sample Request/Response chi2 Example 2

Request:
POST url="http://charm:8080/symetry/rest/c1/projects/dice/explore?metric=chi2"

Body:
{"values":[{"targets":[],"inputAttributes":["d1","d2","d3","d4","d5"],"inputAttributeNames":[],"extraParameters":{}}]}


Response:
{"statusCode":"OK","statusString":"OK","values":{"smlInfo":{"pid":"dicep","isDirty":true,"modelsList":[],"modelTypeList":[],"dsList":["dice5.csv"],"attributeNames":["d1$1","d1$5","d1$6","d1$3","d1$4","d1$2","d2$3","d2$2","d2$5","d2$6","d2$4","d2$1","d3$2","d3$1","d3$6","d3$4","d3$3","d3$5","d4$5","d4$2","d4$3","d4$6","d4$4","d4$1","d5$1","d5$2","d5$3","d5$4","d5$6","d5$5"],"attributeIndexes":[0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29],"attributeTypes":["B","B","B","B","B","B","B","B","B","B","B","B","B","B","B","B","B","B","B","B","B","B","B","B","B","B","B","B","B","B"],"modelAssessments":{},"modelPredictions":{},"params":{},"hash":-1,"categorySeparator":"$","type":"cpu","creationDate":1589219373647,"lastModificationDate":1589219374612,"loaded":true,"persisted":true,"histogramEnabled":false},"KSVDMap":{"values":[{"obs$d1:4":174.0,"obs$d2:3":180.0,"obs$d3:2":179.0,"obs$d4:2":155.0,"obs$d5:1":169.0,"obs$d1:3":156.0,"obs$d2:2":162.0,"obs$d3:1":175.0,"obs$d4:3":179.0,"obs$d5:2":172.0,"obs$d1:2":161.0,"obs$d2:1":145.0,"obs$d4:4":196.0,"obs$d5:3":164.0,"df":20.0,"obs$d1:1":182.0,"obs$d4:5":166.0,"obs$d5:4":177.0,"obs$d3:6":155.0,"obs$d2:6":178.0,"obs$d3:5":174.0,"pval":0.20470000000000002,"obs$d1:6":149.0,"obs$d2:5":141.0,"obs$d3:4":168.0,"obs$d1:5":178.0,"obs$d2:4":194.0,"obs$d3:3":149.0,"obs$d4:1":158.0,"obs$d4:6":146.0,"obs$d5:5":156.0,"obs$d5:6":162.0,"coef":0.0334,"chiStat":24.9148}]}}}

VIF Exploration

This rest endpoint return the variance inflation factor for a list of features. This API function is asynchronous. If it succeeds, it returns a 202 response, along with a Location header that specifies the job URL. For more information about SymetryML asynchronous job please refer the section on Symetry Job.

URL

POST /symetry/rest/{cid}/projects/{pid}/vif [body=ExploreContext]

HTTP Responses

HTTP Status Code	HTTP Status Message	Description
202	OK	Job accepted.
400	BAD REQUEST	Unknown SymetryML project. `{"statusCode":"BAD_REQUEST","statusString":" + Cannot Find SYMETRYML id[r2] for Customer id [c1]","values":{}}`

HTTP Response Entity

The response is a KSVSMap, the key are the input index as specified in the MLContext body and the value is the variance inflation factor for that index.

Sample Request/Response

Request:
POST url="http://charm:8080/symetry/rest/c1/projects/irisTest/vif"

Body:
{"values":[{"targets":[],"inputAttributes":["2","4"]}]}

Response Header:
Location: http://charm:8080/symetry/rest/c1/jobs/4

Response:
{"statusCode":"ACCEPTED","statusString":"Job Created","values":{}}

Job Request:
GET url="http://charm:8080/symetry/rest/c1/jobs/4

Job Response:
KSVSMap

PCA Exploration

This REST endpoint allows for the request of the principal component analysis (PCA) of a set of attributes.

Note: PCA results are limited to a maximum of 100 dimensions by default to avoid excessively large response bodies. This limit can be configured using the pcaNumDimension parameter in the request body (see below), but the maximum number of Eigen values/vectors returned will not exceed 100.

This API function is asynchronous. If it succeeds, it returns a 202 response, along with a Location header that specifies the job URL. For more information about SymetryML asynchronous job please refer the section on Symetry Job.

URL

POST /symetry/rest/{cid}/projects/{pid}/explorepca?explainedCovariance [body=ExploreContext]

Query Parameters

Parameter	Required/Optional?	Description
explainedCovariance	Optional	This parameter controls how many Eigen vectors/values will be returned in term of explained covariance. Default is 1 which correspond to 100% of explained covariance. Note that in order to avoid sending too much large response body, the maximum number of Eigen values/vectors that will be returned is 100.

Request Body

The request body is gonna be a MLContext that contains information about the inputs to use. It can also contains the following extra parameter pcaNumDimension to specify a maximum number of dimension to return. The default value is 100.

HTTP Responses

HTTP Status Code	HTTP Status Message	Description
200	OK	Success.
400	BAD REQUEST	Unknown SymetryML project. `{"statusCode":"BAD_REQUEST","statusString":" + Cannot Find SYMETRYML id[r2] for Customer id [c1]","values":{}}`

HTTP Response Entity

The response contains three keys:

pcaVectors
pcaValues
pcaSumValues

The value for the pcaVectors and pcaValues keys is a Matrix stored in a DataFrame JSON data structure. The pcaSumValues key consists of a number that is the sum of all the Eigen values. See DataFrame for information on the DataFrame JSON structure.

HTTP Response Entity	Example
pcaVectors	A DataFrame containing the PCA vectors.
pcaValues	A DataFrame containing the PCA values.
pcaSumValues	A number representing the total sum of the PCA values. You can use that number to compute the % of covariance explained by a given Eigen vectors.

Sample Request/Response

Request:
POST url="http://charm:8080/symetry/rest/c1/projects/irisTest/explorepca"

Body:
{"values":[{"targets":[],"inputAttributes":["2","4"]}]}

Response Header:
Location: http://charm:8080/symetry/rest/c1/jobs/4

Response:
{"statusCode":"ACCEPTED","statusString":"Job Created","values":{}}

Job Request:
GET url="http://charm:8080/symetry/rest/c1/jobs/4

Job Response:
{pcaVectors, pcaValues, pcaSumValues}

SVD Exploration

This REST endpoint allows to get the singular values decomposition (SVD) of a set of attributes.

The request body needs to contains the list of input attributes id to be used while computing the singular value decomposition. Refer to section ExploreContext for details.

URL

POST /symetry/rest/{cid}/projects/{pid}/exploresvd [body=ExploreContext]

HTTP Responses

HTTP Status Code	HTTP Status Message	Description
200	OK	Success.
400	BAD REQUEST	Unknown SymetryML project. `{"statusCode":"BAD_REQUEST","statusString":" + Cannot Find SYMETRYML id[r2] for Customer id [c1]","values":{}}`

HTTP Response Entity

HTTP Response Entity	Example
KSVSMap	A map<String, String> map containing singular values along their attributes names. Each key/value pair in the result map contains: key = a SVD computed number, that is a singular value value = The name of the attribute

Sample Request/Response

Request:
POST url="http://charm:8080/symetry/rest/c1/projects/irisTest/exploresvd"

Body:
{"values":[{"targets":[],"inputAttributes":["2","4", "3", "5"]}]}

Response Header:
Location: http://charm:8080/symetry/rest/c1/jobs/4

Response:
{"statusCode":"ACCEPTED","statusString":"Job Created","values":{}}

Job Request:
GET url="http://charm:8080/symetry/rest/c1/jobs/4

Job Response:
KSVSMAP{}

SVD Features Selection

This REST endpoint allows to use singular value decomposition to perform feature selection. This endpoint will return a subset of attributes that are not singular based on the SVD algorithms. The SVD algorithms might be used multiple time internally to eventually reach the solution.

The request body needs to contains the list of input attributes id to be used while computing the singular value decomposition. Refer to section ExploreContext for details.

URL

POST /symetry/rest/{cid}/projects/{pid}/exploresvdfs [body=ExploreContext]

HTTP Responses

HTTP Status Code	HTTP Status Message	Description
200	OK	Success.
400	BAD REQUEST	Unknown SymetryML project. `{"statusCode":"BAD_REQUEST","statusString":" + Cannot Find SYMETRYML id[r2] for Customer id [c1]","values":{}}`

HTTP Response Entity

HTTP Response Entity	Example
KSVSMap	A map<String, String> map containing singular values along their attributes names. Each key/value pair in the result map contains: key = a SVD computed number, that is a singular value value = The name of the attribute

Density Estimates

This method returns the density estimate / histogram for any continuous or binary attributes in your SymetryML project. Histogram building must have been enabled - for the wanted symetry project - before invoking this rest endpoint. Histogram building for a project can be enabled by 2 means:

When creating the project with the enableHistogram query parameter. See Create Project endpoint.
Explicitly invoking the Enable Histogram rest endpoint on any given project.

URL

POST /symetry/rest/{cid}/projects/{pid}/densityEstimate [body=ExploreContext]

Query Parameters

Parameter	Required / Optional	Description
useLocal	Optional	If "true", use local project for federated projects

MLContext Parameters for Density Estimates

For each features / attributes for which histogram are requested a MLContext is needed inside the ExplorationContext. The following tables describes the extra parameters can be used:

Parameter	Required / Optional	Description
hist_bins	Required	Number of bins in the histogram
hist_max	Optional	Minimum value of the histogram on the x axis
hist_min	Optional	Maximum value of the histogram on the x axis

HTTP Responses

HTTP Status Code	HTTP Status Message	Description
200	OK	Success.
400	BAD REQUEST	Unknown SymetryML project. `{"statusCode":"BAD_REQUEST","statusString":" + Cannot Find SYMETRYML id[r2] for Customer id [c1]","values":{}}`

HTTP Response Entity

HTTP Response Entity	Example
DensityList	see sample response below

Sample Request/Response

Request:
POST url="http://charm:8080/symetry/rest/c1/projects/irisTest/densityEstimate"

Body:
{"values":[{"inputAttributeNames":["sepal_length"],"inputAttributes":[],"targets":[],"extraParameters":{"hist_bins":"20"}},{"inputAttributeNames":["sepal_width"],"inputAttributes":[],"targets":[],"extraParameters":{"hist_bins":"20"}},{"inputAttributeNames":["petal_length"],"inputAttributes":[],"targets":[],"extraParameters":{"hist_bins":"20"}},{"inputAttributeNames":["petal_width"],"inputAttributes":[],"targets":[],"extraParameters":{"hist_bins":"20"}},{"inputAttributeNames":["sepal_lengt_b1"],"inputAttributes":[],"targets":[],"extraParameters":{"hist_bins":"2"}},{"inputAttributeNames":["sepal_lengt_b2"],"inputAttributes":[],"targets":[],"extraParameters":{"hist_bins":"2"}},{"inputAttributeNames":["sepal_width_b1"],"inputAttributes":[],"targets":[],"extraParameters":{"hist_bins":"2"}},{"inputAttributeNames":["sepal_width_b2"],"inputAttributes":[],"targets":[],"extraParameters":{"hist_bins":"2"}},{"inputAttributeNames":["petal_length_b1"],"inputAttributes":[],"targets":[],"extraParameters":{"hist_bins":"2"}},{"inputAttributeNames":["petal_length_b2"],"inputAttributes":[],"targets":[],"extraParameters":{"hist_bins":"2"}},{"inputAttributeNames":["petal_width_b1"],"inputAttributes":[],"targets":[],"extraParameters":{"hist_bins":"2"}},{"inputAttributeNames":["petal_width_b2"],"inputAttributes":[],"targets":[],"extraParameters":{"hist_bins":"2"}},{"inputAttributeNames":["Iris_setosa"],"inputAttributes":[],"targets":[],"extraParameters":{"hist_bins":"2"}},{"inputAttributeNames":["Iris_versicolor"],"inputAttributes":[],"targets":[],"extraParameters":{"hist_bins":"2"}},{"inputAttributeNames":["Iris_virginica"],"inputAttributes":[],"targets":[],"extraParameters":{"hist_bins":"2"}}]}


Response:
{"statusCode":"OK","statusString":"OK","values":{"densityList":{"values":[{"attrName":"sepal_length","min":4.3,"max":7.9,"width":0.18000000000000002,"histogram":[4.0,5.0,7.0,16.0,9.0,5.0,13.0,14.0,10.0,6.0,10.0,16.0,7.0,11.0,5.0,1.0,4.0,1.0,5.0,1.0]},{"attrName":"sepal_width","min":2.0,"max":4.4,"width":0.12000000000000002,"histogram":[1.0,3.0,4.0,3.0,8.0,14.0,14.0,10.0,26.0,12.0,19.0,12.0,6.0,3.0,9.0,2.0,1.0,1.0,1.0,1.0]},{"attrName":"petal_length","min":1.0,"max":6.9,"width":0.29500000000000004,"histogram":[4.0,33.0,11.0,2.0,0.0,0.0,1.0,2.0,3.0,5.0,12.0,14.0,12.0,17.0,6.0,12.0,7.0,4.0,2.0,3.0]},{"attrName":"petal_width","min":0.1,"max":2.5,"width":0.12,"histogram":[34.0,7.0,7.0,1.0,1.0,0.0,0.0,7.0,3.0,5.0,21.0,12.0,4.0,2.0,17.0,6.0,6.0,3.0,8.0,6.0]},{"attrName":"sepal_lengt_b1","min":0.0,"max":1.0,"width":0.5,"histogram":[67.0,83.0]},{"attrName":"sepal_lengt_b2","min":0.0,"max":1.0,"width":0.5,"histogram":[83.0,67.0]},{"attrName":"sepal_width_b1","min":0.0,"max":1.0,"width":0.5,"histogram":[93.0,57.0]},{"attrName":"sepal_width_b2","min":0.0,"max":1.0,"width":0.5,"histogram":[57.0,93.0]},{"attrName":"petal_length_b1","min":0.0,"max":1.0,"width":0.5,"histogram":[89.0,61.0]},{"attrName":"petal_length_b2","min":0.0,"max":1.0,"width":0.5,"histogram":[61.0,89.0]},{"attrName":"petal_width_b1","min":0.0,"max":1.0,"width":0.5,"histogram":[100.0,50.0]},{"attrName":"petal_width_b2","min":0.0,"max":1.0,"width":0.5,"histogram":[50.0,100.0]},{"attrName":"Iris_setosa","min":0.0,"max":1.0,"width":0.5,"histogram":[100.0,50.0]},{"attrName":"Iris_versicolor","min":0.0,"max":1.0,"width":0.5,"histogram":[100.0,50.0]},{"attrName":"Iris_virginica","min":0.0,"max":1.0,"width":0.5,"histogram":[100.0,50.0]}]}}}

Conditional Density Estimates

This method returns conditional density estimates (histograms) for pairs of attributes in your SymetryML project. It is useful for bivariate analysis with conditional distributions. Histogram building must have been enabled for the project before invoking this endpoint.

URL

POST /symetry/rest/{cid}/projects/{pid}/conditionalDensityEstimate [body=ExploreContext]

Query Parameters

Parameter	Required / Optional	Default	Description
tps	Optional	10	Thread pool size for parallel processing
useLocal	Optional	-	If "true", use local project for federated projects

Request Body

The request body is an ExploreContext containing pairs of attributes. Each MLContext should specify:

The conditional attribute (the attribute to condition on)
The target attribute (the attribute for which to compute the conditional density)

HTTP Responses

HTTP Status Code	HTTP Status Message	Description
200	OK	Success.
400	BAD REQUEST	Unknown SymetryML project. `{"statusCode":"BAD_REQUEST","statusString":"Cannot Find SYMETRYML id[r2] for Customer id [c1]","values":{}}`

HTTP Response Entity

HTTP Response Entity	Example
DensityList	Contains conditional density estimates for each pair of attributes requested.

Sample Request/Response

Request:
POST url="http://charm:8080/symetry/rest/c1/projects/irisTest/conditionalDensityEstimate?tps=10"

Body:
{"values":[{"inputAttributeNames":["sepal_length","petal_length"],"inputAttributes":[],"targets":[],"extraParameters":{"hist_bins":"20"}}]}

Response:
{"statusCode":"OK","statusString":"OK","values":{"densityList":{"values":[{"attrName":"petal_length|sepal_length","min":1.0,"max":6.9,"width":0.295,"histogram":[...]}]}}}

Exploration API

On this page