Skip to content

[ML] Allow configuration of prediction column name in data_frame_analyzer #587

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 3 commits into from
Aug 7, 2019
Merged

[ML] Allow configuration of prediction column name in data_frame_analyzer #587

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
61cf033
Allow configuration of prediction column name and better default
tveasey Aug 6, 2019
9ebb67c
Test fix
tveasey Aug 7, 2019
4a92476
Rename per review
tveasey Aug 7, 2019
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 2 additions & 1 deletion include/api/CDataFrameBoostedTreeRunner.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -54,7 +54,8 @@ class API_EXPORT CDataFrameBoostedTreeRunner final : public CDataFrameAnalysisRu
private:
// Note custom config is written directly to the factory object.

std::string m_DependentVariable;
std::string m_DependentVariableFieldName;
std::string m_PredictionFieldName;
TBoostedTreeFactoryUPtr m_BoostedTreeFactory;
TBoostedTreeUPtr m_BoostedTree;
};
Expand Down
26 changes: 14 additions & 12 deletions lib/api/CDataFrameBoostedTreeRunner.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -23,7 +23,8 @@ namespace ml {
namespace api {
namespace {
// Configuration
const std::string DEPENDENT_VARIABLE{"dependent_variable"};
const std::string DEPENDENT_VARIABLE_NAME{"dependent_variable"};
const std::string PREDICTION_FIELD_NAME{"prediction_field_name"};
const std::string LAMBDA{"lambda"};
const std::string GAMMA{"gamma"};
const std::string ETA{"eta"};
Expand All @@ -32,8 +33,10 @@ const std::string FEATURE_BAG_FRACTION{"feature_bag_fraction"};

const CDataFrameAnalysisConfigReader PARAMETER_READER{[] {
CDataFrameAnalysisConfigReader theReader;
theReader.addParameter(DEPENDENT_VARIABLE,
theReader.addParameter(DEPENDENT_VARIABLE_NAME,
CDataFrameAnalysisConfigReader::E_RequiredParameter);
theReader.addParameter(PREDICTION_FIELD_NAME,
CDataFrameAnalysisConfigReader::E_OptionalParameter);
// TODO objective function, support train and predict.
theReader.addParameter(LAMBDA, CDataFrameAnalysisConfigReader::E_OptionalParameter);
theReader.addParameter(GAMMA, CDataFrameAnalysisConfigReader::E_OptionalParameter);
Expand All @@ -44,9 +47,6 @@ const CDataFrameAnalysisConfigReader PARAMETER_READER{[] {
CDataFrameAnalysisConfigReader::E_OptionalParameter);
return theReader;
}()};

// Output
const std::string PREDICTION{"prediction"};
}

CDataFrameBoostedTreeRunner::CDataFrameBoostedTreeRunner(const CDataFrameAnalysisSpecification& spec,
Expand All @@ -55,7 +55,10 @@ CDataFrameBoostedTreeRunner::CDataFrameBoostedTreeRunner(const CDataFrameAnalysi

auto parameters = PARAMETER_READER.read(jsonParameters);

m_DependentVariable = parameters[DEPENDENT_VARIABLE].as<std::string>();
m_DependentVariableFieldName = parameters[DEPENDENT_VARIABLE_NAME].as<std::string>();

m_PredictionFieldName = parameters[PREDICTION_FIELD_NAME].fallback(
m_DependentVariableFieldName + "_prediction");

std::size_t maximumNumberTrees{
parameters[MAXIMUM_NUMBER_TREES].fallback(std::size_t{0})};
Expand Down Expand Up @@ -117,21 +120,20 @@ void CDataFrameBoostedTreeRunner::writeOneRow(const TStrVec&,
HANDLE_FATAL(<< "Internal error: boosted tree object missing. Please report this error.");
} else {
writer.StartObject();
writer.Key(PREDICTION);
writer.Key(m_PredictionFieldName);
writer.Double(row[m_BoostedTree->columnHoldingPrediction(row.numberColumns())]);
writer.EndObject();
}
}

void CDataFrameBoostedTreeRunner::runImpl(const TStrVec& featureNames,
core::CDataFrame& frame) {
auto dependentVariableColumn =
std::find(featureNames.begin(), featureNames.end(), m_DependentVariable);
auto dependentVariableColumn = std::find(
featureNames.begin(), featureNames.end(), m_DependentVariableFieldName);
if (dependentVariableColumn == featureNames.end()) {
HANDLE_FATAL(<< "Input error: supplied variable to predict '"
<< m_DependentVariable << "' is missing from training data "
<< core::CContainerPrinter::print(featureNames)
<< ". Please report this problem.");
<< m_DependentVariableFieldName << "' is missing from training"
<< " data " << core::CContainerPrinter::print(featureNames));
} else {
m_BoostedTree = m_BoostedTreeFactory->buildFor(
frame, dependentVariableColumn - featureNames.begin());
Expand Down
22 changes: 11 additions & 11 deletions lib/api/unittest/CDataFrameAnalyzerTest.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -117,7 +117,7 @@ regressionSpec(std::string dependentVariable,
rows, 5, memoryLimit, 1, categoricalFieldNames, true,
test::CTestTmpDir::tmpDir(), "ml", "regression", parameters)};

LOG_DEBUG(<< "spec =\n" << spec);
LOG_TRACE(<< "spec =\n" << spec);

return std::make_unique<api::CDataFrameAnalysisSpecification>(spec);
}
Expand Down Expand Up @@ -294,7 +294,7 @@ void CDataFrameAnalyzerTest::testWithoutControlMessages() {
analyzer.run();

rapidjson::Document results;
rapidjson::ParseResult ok(results.Parse(output.str().c_str()));
rapidjson::ParseResult ok(results.Parse(output.str()));
CPPUNIT_ASSERT(static_cast<bool>(ok) == true);

auto expectedScore = expectedScores.begin();
Expand Down Expand Up @@ -340,7 +340,7 @@ void CDataFrameAnalyzerTest::testRunOutlierDetection() {
analyzer.handleRecord(fieldNames, {"", "", "", "", "", "", "$"});

rapidjson::Document results;
rapidjson::ParseResult ok(results.Parse(output.str().c_str()));
rapidjson::ParseResult ok(results.Parse(output.str()));
CPPUNIT_ASSERT(static_cast<bool>(ok) == true);

auto expectedScore = expectedScores.begin();
Expand Down Expand Up @@ -394,7 +394,7 @@ void CDataFrameAnalyzerTest::testRunOutlierDetectionPartitioned() {
analyzer.handleRecord(fieldNames, {"", "", "", "", "", "", "$"});

rapidjson::Document results;
rapidjson::ParseResult ok(results.Parse(output.str().c_str()));
rapidjson::ParseResult ok(results.Parse(output.str()));
CPPUNIT_ASSERT(static_cast<bool>(ok) == true);

auto expectedScore = expectedScores.begin();
Expand Down Expand Up @@ -441,7 +441,7 @@ void CDataFrameAnalyzerTest::testRunOutlierFeatureInfluences() {
analyzer.handleRecord(fieldNames, {"", "", "", "", "", "", "$"});

rapidjson::Document results;
rapidjson::ParseResult ok(results.Parse(output.str().c_str()));
rapidjson::ParseResult ok(results.Parse(output.str()));
CPPUNIT_ASSERT(static_cast<bool>(ok) == true);

auto expectedFeatureInfluence = expectedFeatureInfluences.begin();
Expand Down Expand Up @@ -492,7 +492,7 @@ void CDataFrameAnalyzerTest::testRunOutlierDetectionWithParams() {
analyzer.handleRecord(fieldNames, {"", "", "", "", "", "", "$"});

rapidjson::Document results;
rapidjson::ParseResult ok(results.Parse(output.str().c_str()));
rapidjson::ParseResult ok(results.Parse(output.str()));
CPPUNIT_ASSERT(static_cast<bool>(ok) == true);

auto expectedScore = expectedScores.begin();
Expand Down Expand Up @@ -530,7 +530,7 @@ void CDataFrameAnalyzerTest::testRunBoostedTreeTraining() {
analyzer.handleRecord(fieldNames, {"", "", "", "", "", "", "$"});

rapidjson::Document results;
rapidjson::ParseResult ok(results.Parse(output.str().c_str()));
rapidjson::ParseResult ok(results.Parse(output.str()));
CPPUNIT_ASSERT(static_cast<bool>(ok) == true);

auto expectedPrediction = expectedPredictions.begin();
Expand All @@ -540,7 +540,7 @@ void CDataFrameAnalyzerTest::testRunBoostedTreeTraining() {
CPPUNIT_ASSERT(expectedPrediction != expectedPredictions.end());
CPPUNIT_ASSERT_DOUBLES_EQUAL(
*expectedPrediction,
result["row_results"]["results"]["ml"]["prediction"].GetDouble(),
result["row_results"]["results"]["ml"]["c5_prediction"].GetDouble(),
1e-4 * std::fabs(*expectedPrediction));
++expectedPrediction;
CPPUNIT_ASSERT(result.HasMember("progress_percent") == false);
Expand Down Expand Up @@ -584,7 +584,7 @@ void CDataFrameAnalyzerTest::testRunBoostedTreeTrainingWithParams() {
analyzer.handleRecord(fieldNames, {"", "", "", "", "", "", "$"});

rapidjson::Document results;
rapidjson::ParseResult ok(results.Parse(output.str().c_str()));
rapidjson::ParseResult ok(results.Parse(output.str()));
CPPUNIT_ASSERT(static_cast<bool>(ok) == true);

auto expectedPrediction = expectedPredictions.begin();
Expand All @@ -594,7 +594,7 @@ void CDataFrameAnalyzerTest::testRunBoostedTreeTrainingWithParams() {
CPPUNIT_ASSERT(expectedPrediction != expectedPredictions.end());
CPPUNIT_ASSERT_DOUBLES_EQUAL(
*expectedPrediction,
result["row_results"]["results"]["ml"]["prediction"].GetDouble(),
result["row_results"]["results"]["ml"]["c5_prediction"].GetDouble(),
1e-4 * std::fabs(*expectedPrediction));
++expectedPrediction;
CPPUNIT_ASSERT(result.HasMember("progress_percent") == false);
Expand Down Expand Up @@ -748,7 +748,7 @@ void CDataFrameAnalyzerTest::testRoundTripDocHashes() {
{"", "", "", "", "", "", "$"});

rapidjson::Document results;
rapidjson::ParseResult ok(results.Parse(output.str().c_str()));
rapidjson::ParseResult ok(results.Parse(output.str()));
CPPUNIT_ASSERT(static_cast<bool>(ok) == true);

int expectedHash{0};
Expand Down