Skip to content
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

[SPARK-11215][ML] Add multiple columns support to StringIndexer #20146

Closed
wants to merge 28 commits into from
Closed

[SPARK-11215][ML] Add multiple columns support to StringIndexer #20146

Show file tree
Hide file tree
Changes from 3 commits
Commits
Show all changes
28 commits
Select commit Hold shift + click to select a range
bb990f1
Add multi-column support to StringIndexer.
viirya Jan 3, 2018
26cc94b
Fix glm test.
viirya Jan 7, 2018
540c364
Merge remote-tracking branch 'upstream/master' into SPARK-11215
viirya Jan 16, 2018
18acbbf
Improve test cases.
viirya Jan 24, 2018
b884fb5
Merge remote-tracking branch 'upstream/master' into SPARK-11215
viirya Jan 27, 2018
76ff7bf
Merge remote-tracking branch 'upstream/master' into SPARK-11215
viirya Apr 19, 2018
50af02e
Change from 2.3 to 2.4.
viirya Apr 19, 2018
c1be2c7
Address comments.
viirya Apr 23, 2018
ed35d87
Address comment.
viirya Apr 23, 2018
a1dcfda
Use SQL Aggregator for counting string labels.
viirya Apr 24, 2018
c168522
Merge remote-tracking branch 'upstream/master' into SPARK-11215
viirya Apr 25, 2018
a6551b0
Drop NA values for both frequency and alphabet order types.
viirya Apr 26, 2018
c003bd3
Merge remote-tracking branch 'upstream/master' into SPARK-11215
viirya Jul 18, 2018
ea3a46b
Merge remote-tracking branch 'upstream/master' into SPARK-11215
viirya Dec 8, 2018
f7102e9
Address comments.
viirya Dec 8, 2018
301fa4c
Update ml document.
viirya Dec 8, 2018
196db63
Address comments.
viirya Dec 21, 2018
cd1eda0
Add a comment.
viirya Dec 21, 2018
7155c63
Merge remote-tracking branch 'upstream/master' into SPARK-11215
viirya Dec 22, 2018
70009a5
Address part of comments.
viirya Dec 22, 2018
3c6ffc7
Fix null and NaN issue.
viirya Dec 22, 2018
b6ad1e4
Add test for loading model prior to Spark 3.0.
viirya Dec 23, 2018
d6fed35
Address comment.
viirya Dec 30, 2018
0137d67
Unpersist if input is not originally cached. Add deprecated info.
viirya Jan 7, 2019
7a5be12
Revert classification doctests change.
viirya Jan 7, 2019
b33556b
Fix style. Revert deprecated methods.
viirya Jan 12, 2019
c3d2522
Merge remote-tracking branch 'upstream/master' into SPARK-11215
viirya Jan 25, 2019
867e001
Address comments.
viirya Jan 25, 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
6 changes: 3 additions & 3 deletions R/pkg/tests/fulltests/test_mllib_classification.R
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -313,7 +313,7 @@ test_that("spark.mlp", {
# Test predict method
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Actually I think we may remove this test Test predict method. Seems to me, with the tol = 0.5, the prediction may not be very meaningful.

mlpTestDF <- df
mlpPredictions <- collect(select(predict(model, mlpTestDF), "prediction"))
expect_equal(head(mlpPredictions$prediction, 6), c("1.0", "0.0", "0.0", "0.0", "0.0", "0.0"))
expect_equal(head(mlpPredictions$prediction, 6), c("0.0", "1.0", "1.0", "1.0", "1.0", "1.0"))
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is due to the change of how we sort string labels with same frequency under the setting of frequencyDesc/Asc.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I checked the predictions. All 0.0 -> 1.0, and all1.0 -> 0.0.


# Test model save/load
if (windows_with_hadoop()) {
Expand Down Expand Up @@ -348,12 +348,12 @@ test_that("spark.mlp", {

# Test random seed
# default seed
model <- spark.mlp(df, label ~ features, layers = c(4, 5, 4, 3), maxIter = 10)
model <- spark.mlp(df, label ~ features, layers = c(4, 5, 4, 3), maxIter = 100)
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Seems maxIter = 10 is not stable. I increased to 100 to stabilize the predictions.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

can you check if the run time increases significantly? this is an issue before - see SPARK-21693

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more. 

> start.time <- Sys.time()
> model <- spark.mlp(df, label ~ features, layers = c(4, 5, 4, 3), maxIter = 10)
> end.time <- Sys.time()
> time.taken <- end.time - start.time
> time.taken
Time difference of 1.780564 secs
> start.time <- Sys.time()
> model <- spark.mlp(df, label ~ features, layers = c(4, 5, 4, 3), maxIter = 100)
> end.time <- Sys.time()
> time.taken <- end.time - start.time
> time.taken
Time difference of 5.728089 secs

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Ahh, @viirya, would you mind if I ask to check it after setting spark.sparkr.use.daemon to false too?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Ran it again with the config:

> sparkR.conf("spark.sparkr.use.daemon")                                                                         
$spark.sparkr.use.daemon
[1] "false"
> start.time <- Sys.time()
> model <- spark.mlp(df, label ~ features, layers = c(4, 5, 4, 3), maxIter = 10)
> end.time <- Sys.time()
> time.taken <- end.time - start.time
> time.taken
Time difference of 1.704288 secs
> start.time <- Sys.time()
> model <- spark.mlp(df, label ~ features, layers = c(4, 5, 4, 3), maxIter = 100)
> end.time <- Sys.time()
> time.taken <- end.time - start.time
> time.taken
Time difference of 5.135418 secs

mlpPredictions <- collect(select(predict(model, mlpTestDF), "prediction"))
expect_equal(head(mlpPredictions$prediction, 10),
c("1.0", "1.0", "1.0", "1.0", "0.0", "1.0", "2.0", "2.0", "1.0", "0.0"))
# seed equals 10
model <- spark.mlp(df, label ~ features, layers = c(4, 5, 4, 3), maxIter = 10, seed = 10)
model <- spark.mlp(df, label ~ features, layers = c(4, 5, 4, 3), maxIter = 100, seed = 10)
mlpPredictions <- collect(select(predict(model, mlpTestDF), "prediction"))
expect_equal(head(mlpPredictions$prediction, 10),
c("1.0", "1.0", "1.0", "1.0", "0.0", "1.0", "2.0", "2.0", "1.0", "0.0"))
Expand Down
42 changes: 29 additions & 13 deletions R/pkg/tests/fulltests/test_mllib_regression.R
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -102,10 +102,18 @@ test_that("spark.glm and predict", {
})

test_that("spark.glm summary", {
# prepare dataset
Sepal.Length <- c(2.0, 1.5, 1.8, 3.4, 5.1, 1.8, 1.0, 2.3)
Sepal.Width <- c(2.1, 2.3, 5.4, 4.7, 3.1, 2.1, 3.1, 5.5)
Petal.Length <- c(1.8, 2.1, 7.1, 2.5, 3.7, 6.3, 2.2, 7.2)
Species <- c("setosa", "versicolor", "versicolor", "versicolor", "virginica", "virginica",
"versicolor", "virginica")
dataset <- data.frame(Sepal.Length, Sepal.Width, Petal.Length, Species, stringsAsFactors = TRUE)

# gaussian family
training <- suppressWarnings(createDataFrame(iris))
training <- suppressWarnings(createDataFrame(dataset))
stats <- summary(spark.glm(training, Sepal_Width ~ Sepal_Length + Species))
rStats <- summary(glm(Sepal.Width ~ Sepal.Length + Species, data = iris))
rStats <- summary(glm(Sepal.Width ~ Sepal.Length + Species, data = dataset))

# test summary coefficients return matrix type
expect_true(class(stats$coefficients) == "matrix")
Expand All @@ -126,15 +134,15 @@ test_that("spark.glm summary", {

out <- capture.output(print(stats))
expect_match(out[2], "Deviance Residuals:")
expect_true(any(grepl("AIC: 59.22", out)))
expect_true(any(grepl("AIC: 35.84", out)))
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

R glm's AIC: 35.839:

> out <- capture.output(print(rStats))
> out
 [1] ""                                                                        
 [2] "Call:"                                                                   
 [3] "glm(formula = Sepal.Width ~ Sepal.Length + Species, data = dataset)"     
 [4] ""                                                                        
 [5] "Deviance Residuals: "                                                    
 [6] "      1        2        3        4        5        6        7        8  "
 [7] " 0.0000  -1.4932   1.5491   0.5411  -0.8581  -1.2228  -0.5969   2.0809  "
 [8] ""                                                                        
 [9] "Coefficients:"                                                           
[10] "                  Estimate Std. Error t value Pr(>|t|)"                  
[11] "(Intercept)         1.7150     2.0492   0.837    0.450"                  
[12] "Sepal.Length        0.1925     0.5566   0.346    0.747"                  
[13] "Speciesversicolor   1.7894     1.9240   0.930    0.405"                  
[14] "Speciesvirginica    1.2613     2.0735   0.608    0.576"                  
[15] ""                                                                        
[16] "(Dispersion parameter for gaussian family taken to be 2.960032)"         
[17] ""                                                                        
[18] "    Null deviance: 14.719  on 7  degrees of freedom"                     
[19] "Residual deviance: 11.840  on 4  degrees of freedom"                     
[20] "AIC: 35.839"                                                             
[21] ""                                                                        
[22] "Number of Fisher Scoring iterations: 2"                                  
[23] ""


# binomial family
df <- suppressWarnings(createDataFrame(iris))
df <- suppressWarnings(createDataFrame(dataset))
training <- df[df$Species %in% c("versicolor", "virginica"), ]
stats <- summary(spark.glm(training, Species ~ Sepal_Length + Sepal_Width,
family = binomial(link = "logit")))

rTraining <- iris[iris$Species %in% c("versicolor", "virginica"), ]
rTraining <- dataset[dataset$Species %in% c("versicolor", "virginica"), ]
rStats <- summary(glm(Species ~ Sepal.Length + Sepal.Width, data = rTraining,
family = binomial(link = "logit")))

Expand Down Expand Up @@ -174,17 +182,17 @@ test_that("spark.glm summary", {
expect_equal(stats$aic, rStats$aic)

# Test spark.glm works with offset
training <- suppressWarnings(createDataFrame(iris))
training <- suppressWarnings(createDataFrame(dataset))
stats <- summary(spark.glm(training, Sepal_Width ~ Sepal_Length + Species,
family = poisson(), offsetCol = "Petal_Length"))
rStats <- suppressWarnings(summary(glm(Sepal.Width ~ Sepal.Length + Species,
data = iris, family = poisson(), offset = iris$Petal.Length)))
data = dataset, family = poisson(), offset = dataset$Petal.Length)))
expect_true(all(abs(rStats$coefficients - stats$coefficients) < 1e-3))

# Test summary works on base GLM models
baseModel <- stats::glm(Sepal.Width ~ Sepal.Length + Species, data = iris)
baseModel <- stats::glm(Sepal.Width ~ Sepal.Length + Species, data = dataset)
baseSummary <- summary(baseModel)
expect_true(abs(baseSummary$deviance - 12.19313) < 1e-4)
expect_true(abs(baseSummary$deviance - 11.84013) < 1e-4)
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

R glm:

> baseSummary <- summary(stats::glm(Sepal.Width ~ Sepal.Length + Species, data = dataset))
> baseSummary$deviance
[1] 11.84013

Spark glm:

> baseSummary <- summary(spark.glm(training, Sepal_Width ~ Sepal_Length + Species))
> baseSummary$deviance
[1] 11.84013


# Test spark.glm works with regularization parameter
data <- as.data.frame(cbind(a1, a2, b))
Expand Down Expand Up @@ -300,11 +308,19 @@ test_that("glm and predict", {
})

test_that("glm summary", {
# prepare dataset
Sepal.Length <- c(2.0, 1.5, 1.8, 3.4, 5.1, 1.8, 1.0, 2.3)
Sepal.Width <- c(2.1, 2.3, 5.4, 4.7, 3.1, 2.1, 3.1, 5.5)
Petal.Length <- c(1.8, 2.1, 7.1, 2.5, 3.7, 6.3, 2.2, 7.2)
Species <- c("setosa", "versicolor", "versicolor", "versicolor", "virginica", "virginica",
"versicolor", "virginica")
dataset <- data.frame(Sepal.Length, Sepal.Width, Petal.Length, Species, stringsAsFactors = TRUE)

# gaussian family
training <- suppressWarnings(createDataFrame(iris))
training <- suppressWarnings(createDataFrame(dataset))
stats <- summary(glm(Sepal_Width ~ Sepal_Length + Species, data = training))

rStats <- summary(glm(Sepal.Width ~ Sepal.Length + Species, data = iris))
rStats <- summary(glm(Sepal.Width ~ Sepal.Length + Species, data = dataset))

coefs <- stats$coefficients
rCoefs <- rStats$coefficients
Expand All @@ -320,12 +336,12 @@ test_that("glm summary", {
expect_equal(stats$aic, rStats$aic)

# binomial family
df <- suppressWarnings(createDataFrame(iris))
df <- suppressWarnings(createDataFrame(dataset))
training <- df[df$Species %in% c("versicolor", "virginica"), ]
stats <- summary(glm(Species ~ Sepal_Length + Sepal_Width, data = training,
family = binomial(link = "logit")))

rTraining <- iris[iris$Species %in% c("versicolor", "virginica"), ]
rTraining <- dataset[dataset$Species %in% c("versicolor", "virginica"), ]
rStats <- summary(glm(Species ~ Sepal.Length + Sepal.Width, data = rTraining,
family = binomial(link = "logit")))

Expand Down
Loading