You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@systemml.apache.org by GitBox <gi...@apache.org> on 2020/06/01 03:54:01 UTC
[GitHub] [systemml] t201 opened a new pull request #933: KMeans
t201 opened a new pull request #933:
URL: https://github.com/apache/systemml/pull/933
[KMeans.txt](https://github.com/apache/systemml/files/4709296/KMeans.txt)
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [systemml] j143 removed a comment on pull request #933: KMeans
Posted by GitBox <gi...@apache.org>.
j143 removed a comment on pull request #933:
URL: https://github.com/apache/systemml/pull/933#issuecomment-636880414
@t201 - I will review this tonight. Thank you.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [systemml] j143 commented on a change in pull request #933: [DOC] Documentation for builtin KMeans function
Posted by GitBox <gi...@apache.org>.
j143 commented on a change in pull request #933:
URL: https://github.com/apache/systemml/pull/933#discussion_r434760157
##########
File path: docs/builtins-reference.md
##########
@@ -143,12 +139,13 @@ kmeans(x, centers, iter.max, nstart)
| String |The output matrix with the centroids |
+
+**DML-bodied built-in functions** are written as DML-Scripts and executed as such when called.
### Example
-KMeans = rand(Matrix x, centers k=0,int iter.max=25,int nstart=10)
-return(Matrix[numerical] C,Matrix[numerical] Y)
+```h2o.kmeans(x = predictors, k = 100, estimate_k = T, standardize = F,
Review comment:
The snippet, I've provided above is a DML code to run. 🙂
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [systemml] j143 commented on a change in pull request #933: [DOC] Documentation for builtin KMeans function
Posted by GitBox <gi...@apache.org>.
j143 commented on a change in pull request #933:
URL: https://github.com/apache/systemml/pull/933#discussion_r434346924
##########
File path: docs/builtins-reference.md
##########
@@ -143,12 +139,13 @@ kmeans(x, centers, iter.max, nstart)
| String |The output matrix with the centroids |
+
+**DML-bodied built-in functions** are written as DML-Scripts and executed as such when called.
### Example
-KMeans = rand(Matrix x, centers k=0,int iter.max=25,int nstart=10)
-return(Matrix[numerical] C,Matrix[numerical] Y)
+```h2o.kmeans(x = predictors, k = 100, estimate_k = T, standardize = F,
Review comment:
Are you able to run this code in the colab notebook provided to you, if yes screenshot please.
Also, this seems to be a copy-paste from `H2O` software, different from `SystemDS`.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [systemml] t201 commented on a change in pull request #933: [DOC] Documentation for builtin KMeans function
Posted by GitBox <gi...@apache.org>.
t201 commented on a change in pull request #933:
URL: https://github.com/apache/systemml/pull/933#discussion_r434699150
##########
File path: docs/builtins-reference.md
##########
@@ -143,12 +139,13 @@ kmeans(x, centers, iter.max, nstart)
| String |The output matrix with the centroids |
+
+**DML-bodied built-in functions** are written as DML-Scripts and executed as such when called.
### Example
-KMeans = rand(Matrix x, centers k=0,int iter.max=25,int nstart=10)
-return(Matrix[numerical] C,Matrix[numerical] Y)
+```h2o.kmeans(x = predictors, k = 100, estimate_k = T, standardize = F,
Review comment:
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [systemml] j143 commented on a change in pull request #933: [DOC] Documentation for builtin KMeans function
Posted by GitBox <gi...@apache.org>.
j143 commented on a change in pull request #933:
URL: https://github.com/apache/systemml/pull/933#discussion_r434360738
##########
File path: docs/builtins-reference.md
##########
@@ -143,12 +139,13 @@ kmeans(x, centers, iter.max, nstart)
| String |The output matrix with the centroids |
+
+**DML-bodied built-in functions** are written as DML-Scripts and executed as such when called.
### Example
-KMeans = rand(Matrix x, centers k=0,int iter.max=25,int nstart=10)
-return(Matrix[numerical] C,Matrix[numerical] Y)
+```h2o.kmeans(x = predictors, k = 100, estimate_k = T, standardize = F,
Review comment:
We need to closely understand the this function:
https://github.com/apache/systemml/blob/faec15810e18fc497e950b830811a919f5aa5c5f/scripts/builtin/kmeans.dml#L46
Think about the inputs required and initialize the required arguments.
We can take already implemented functions as examples.
This task is tough and iterative, copy-paste solutions are not available. Some people spend whole days to find the correct combinations, though. 🙂
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [systemml] j143 commented on a change in pull request #933: [DOC] Documentation for builtin KMeans function
Posted by GitBox <gi...@apache.org>.
j143 commented on a change in pull request #933:
URL: https://github.com/apache/systemml/pull/933#discussion_r434292865
##########
File path: docs/builtins-reference.md
##########
@@ -105,9 +106,51 @@ print(toString(D))
Note that reshape construction is not yet supported for **SPARK** execution.
# DML-Bodied Built-In Functions
+# Introduction
+
+
+The DML (Declarative Machine Learning) language has built-in functions which enable access to both low- and high-level function
+
+
+
+## `KMeans`-Function
+
+The kmeans() function in R requires, at a minimum, numeric data and a number of centers (or clusters).
+The cluster centers are pulled out by using $centers.
+
+
+
+### Usage
+```r
+kmeans(x, centers, iter.max, nstart)
+```
+
+### Arguments
+
+| Name | Type | Default | Description |
+| :---- | :------------- | -------- | :------------------------------- |
+| x | Matrix[Numeric] | required |The input Matrix to do KMeans on..|
+|centers | Int | -------- |The no. of cenetrs |
+|iter.max|Int |10 |Max no. of iterations allowed |
+|nstart |Int |10 |No. of random starting positions |
+
+
+
+### Returns
+| Type | Description |
+| :------------- | :---------- |
+| String | The mapping of records to centroids |
+| String |The output matrix with the centroids |
+
+
+### Example
+
+KMeans = rand(Matrix x, centers k=0,int iter.max=25,int nstart=10)
+return(Matrix[numerical] C,Matrix[numerical] Y)
Review comment:
1. Enclose this snippet in triple tick (```)
##########
File path: docs/builtins-reference.md
##########
@@ -105,9 +106,51 @@ print(toString(D))
Note that reshape construction is not yet supported for **SPARK** execution.
# DML-Bodied Built-In Functions
+# Introduction
+
+
+The DML (Declarative Machine Learning) language has built-in functions which enable access to both low- and high-level function
+
+
+
+## `KMeans`-Function
+
+The kmeans() function in R requires, at a minimum, numeric data and a number of centers (or clusters).
+The cluster centers are pulled out by using $centers.
+
+
+
+### Usage
+```r
+kmeans(x, centers, iter.max, nstart)
+```
+
+### Arguments
+
+| Name | Type | Default | Description |
+| :---- | :------------- | -------- | :------------------------------- |
+| x | Matrix[Numeric] | required |The input Matrix to do KMeans on..|
+|centers | Int | -------- |The no. of cenetrs |
+|iter.max|Int |10 |Max no. of iterations allowed |
+|nstart |Int |10 |No. of random starting positions |
+
+
+
+### Returns
+| Type | Description |
+| :------------- | :---------- |
+| String | The mapping of records to centroids |
+| String |The output matrix with the centroids |
+
+
+### Example
+
+KMeans = rand(Matrix x, centers k=0,int iter.max=25,int nstart=10)
+return(Matrix[numerical] C,Matrix[numerical] Y)
Review comment:
2. Example in this case means `a runnable code snippet`.
**For Example:** In the case of `slicefinder` function
We can copy paste this usage code into the Colab and run it!
Which would give the following output:
##########
File path: docs/builtins-reference.md
##########
@@ -105,9 +106,51 @@ print(toString(D))
Note that reshape construction is not yet supported for **SPARK** execution.
# DML-Bodied Built-In Functions
+# Introduction
+
+
+The DML (Declarative Machine Learning) language has built-in functions which enable access to both low- and high-level function
+
+
+
+## `KMeans`-Function
+
+The kmeans() function in R requires, at a minimum, numeric data and a number of centers (or clusters).
+The cluster centers are pulled out by using $centers.
+
+
+
+### Usage
+```r
+kmeans(x, centers, iter.max, nstart)
+```
+
+### Arguments
+
+| Name | Type | Default | Description |
+| :---- | :------------- | -------- | :------------------------------- |
+| x | Matrix[Numeric] | required |The input Matrix to do KMeans on..|
+|centers | Int | -------- |The no. of cenetrs |
+|iter.max|Int |10 |Max no. of iterations allowed |
+|nstart |Int |10 |No. of random starting positions |
Review comment:
1. Consistency to be preferred, such as `| centers |` instead of `|centers |`.
2. In `centers` we can replace `--` instead of --------, in Default column
3. For values in the `Default` column let us prettify the formatting by enclosing the 10 in backtick (`)
##########
File path: docs/builtins-reference.md
##########
@@ -105,9 +106,51 @@ print(toString(D))
Note that reshape construction is not yet supported for **SPARK** execution.
# DML-Bodied Built-In Functions
+# Introduction
+
+
+The DML (Declarative Machine Learning) language has built-in functions which enable access to both low- and high-level function
+
+
+
Review comment:
These lines are not required!
##########
File path: docs/builtins-reference.md
##########
@@ -105,9 +106,51 @@ print(toString(D))
Note that reshape construction is not yet supported for **SPARK** execution.
# DML-Bodied Built-In Functions
+# Introduction
+
+
+The DML (Declarative Machine Learning) language has built-in functions which enable access to both low- and high-level function
+
+
+
+## `KMeans`-Function
+
+The kmeans() function in R requires, at a minimum, numeric data and a number of centers (or clusters).
+The cluster centers are pulled out by using $centers.
+
+
+
+### Usage
+```r
+kmeans(x, centers, iter.max, nstart)
+```
+
+### Arguments
+
+| Name | Type | Default | Description |
+| :---- | :------------- | -------- | :------------------------------- |
+| x | Matrix[Numeric] | required |The input Matrix to do KMeans on..|
+|centers | Int | -------- |The no. of cenetrs |
+|iter.max|Int |10 |Max no. of iterations allowed |
+|nstart |Int |10 |No. of random starting positions |
+
+
+
+### Returns
+| Type | Description |
+| :------------- | :---------- |
+| String | The mapping of records to centroids |
+| String |The output matrix with the centroids |
+
+
+### Example
+
+KMeans = rand(Matrix x, centers k=0,int iter.max=25,int nstart=10)
+return(Matrix[numerical] C,Matrix[numerical] Y)
**DML-bodied built-in functions** are written as DML-Scripts and executed as such when called.
Review comment:
This line comes immediately after `# DML-Bodied Built-In Functions`
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [systemml] j143 commented on pull request #933: [DOC] Documentation for builtin KMeans function
Posted by GitBox <gi...@apache.org>.
j143 commented on pull request #933:
URL: https://github.com/apache/systemml/pull/933#issuecomment-637998718
@nurattarun - Refer https://github.com/apache/systemml/pull/933#discussion_r434296676
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [systemml] t201 commented on a change in pull request #933: [DOC] Documentation for builtin KMeans function
Posted by GitBox <gi...@apache.org>.
t201 commented on a change in pull request #933:
URL: https://github.com/apache/systemml/pull/933#discussion_r434353641
##########
File path: docs/builtins-reference.md
##########
@@ -143,12 +139,13 @@ kmeans(x, centers, iter.max, nstart)
| String |The output matrix with the centroids |
+
+**DML-bodied built-in functions** are written as DML-Scripts and executed as such when called.
### Example
-KMeans = rand(Matrix x, centers k=0,int iter.max=25,int nstart=10)
-return(Matrix[numerical] C,Matrix[numerical] Y)
+```h2o.kmeans(x = predictors, k = 100, estimate_k = T, standardize = F,
Review comment:
I'm unable to find a small code snippet for Kmeans function in R which runs without any error on Collab. I tried with a lot of examples.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [systemml] t201 commented on a change in pull request #933: [DOC] Documentation for builtin KMeans function
Posted by GitBox <gi...@apache.org>.
t201 commented on a change in pull request #933:
URL: https://github.com/apache/systemml/pull/933#discussion_r434353641
##########
File path: docs/builtins-reference.md
##########
@@ -143,12 +139,13 @@ kmeans(x, centers, iter.max, nstart)
| String |The output matrix with the centroids |
+
+**DML-bodied built-in functions** are written as DML-Scripts and executed as such when called.
### Example
-KMeans = rand(Matrix x, centers k=0,int iter.max=25,int nstart=10)
-return(Matrix[numerical] C,Matrix[numerical] Y)
+```h2o.kmeans(x = predictors, k = 100, estimate_k = T, standardize = F,
Review comment:
I'm unable to find a small code snippet for Kmeans function in R which runs without any error on Collab. I tried with a lot of examples. Most of the times there is a syntax error.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [systemml] j143 commented on pull request #933: KMeans
Posted by GitBox <gi...@apache.org>.
j143 commented on pull request #933:
URL: https://github.com/apache/systemml/pull/933#issuecomment-637941282
Hi @t201 - Thanks for working on this function. This is good starting point and there are some minor changes related to formatting and example.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [systemml] j143 commented on pull request #933: KMeans
Posted by GitBox <gi...@apache.org>.
j143 commented on pull request #933:
URL: https://github.com/apache/systemml/pull/933#issuecomment-636880414
@t201 - I will review this tonight. Thank you.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [systemml] asfgit closed pull request #933: [DOC] Documentation for builtin KMeans function
Posted by GitBox <gi...@apache.org>.
asfgit closed pull request #933:
URL: https://github.com/apache/systemml/pull/933
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [systemml] j143 commented on pull request #933: KMeans
Posted by GitBox <gi...@apache.org>.
j143 commented on pull request #933:
URL: https://github.com/apache/systemml/pull/933#issuecomment-636632533
We need to add code like this: https://github.com/apache/systemml/commit/8d020fb9f807750b07d08398caa2d433305819b6
:)
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [systemml] j143 commented on a change in pull request #933: [DOC] Documentation for builtin KMeans function
Posted by GitBox <gi...@apache.org>.
j143 commented on a change in pull request #933:
URL: https://github.com/apache/systemml/pull/933#discussion_r434685969
##########
File path: docs/builtins-reference.md
##########
@@ -143,12 +139,13 @@ kmeans(x, centers, iter.max, nstart)
| String |The output matrix with the centroids |
+
+**DML-bodied built-in functions** are written as DML-Scripts and executed as such when called.
### Example
-KMeans = rand(Matrix x, centers k=0,int iter.max=25,int nstart=10)
-return(Matrix[numerical] C,Matrix[numerical] Y)
+```h2o.kmeans(x = predictors, k = 100, estimate_k = T, standardize = F,
Review comment:
Thanks a lot for persistent work on this, @t201
The following snippet would help as a starting point.
```
X = rand (rows = 10, cols = 10)
kmeans(X = X, k = 0, runs = 10, max_iter = 1000, is_verbose = FALSE, avg_sample_size_per_centroid = 50)
```
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [systemml] j143 commented on a change in pull request #933: [DOC] Documentation for builtin KMeans function
Posted by GitBox <gi...@apache.org>.
j143 commented on a change in pull request #933:
URL: https://github.com/apache/systemml/pull/933#discussion_r434685969
##########
File path: docs/builtins-reference.md
##########
@@ -143,12 +139,13 @@ kmeans(x, centers, iter.max, nstart)
| String |The output matrix with the centroids |
+
+**DML-bodied built-in functions** are written as DML-Scripts and executed as such when called.
### Example
-KMeans = rand(Matrix x, centers k=0,int iter.max=25,int nstart=10)
-return(Matrix[numerical] C,Matrix[numerical] Y)
+```h2o.kmeans(x = predictors, k = 100, estimate_k = T, standardize = F,
Review comment:
```
X = rand (rows = 10, cols = 10)
kmeans(X = X, k = 0, runs = 10, max_iter = 1000, )
```
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [systemml] j143 commented on a change in pull request #933: [DOC] Documentation for builtin KMeans function
Posted by GitBox <gi...@apache.org>.
j143 commented on a change in pull request #933:
URL: https://github.com/apache/systemml/pull/933#discussion_r434346081
##########
File path: docs/builtins-reference.md
##########
@@ -127,12 +123,12 @@ kmeans(x, centers, iter.max, nstart)
### Arguments
-| Name | Type | Default | Description |
-| :---- | :------------- | -------- | :------------------------------- |
-| x | Matrix[Numeric] | required |The input Matrix to do KMeans on..|
-|centers | Int | -------- |The no. of cenetrs |
-|iter.max|Int |10 |Max no. of iterations allowed |
-|nstart |Int |10 |No. of random starting positions |
+| Name | Type | Default | Description |
+| :-- | :-- | -- | :-- |
Review comment:
Can you revert changes of this line. I was only talking about `centers` corresponding!
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [systemml] t201 commented on a change in pull request #933: [DOC] Documentation for builtin KMeans function
Posted by GitBox <gi...@apache.org>.
t201 commented on a change in pull request #933:
URL: https://github.com/apache/systemml/pull/933#discussion_r434352595
##########
File path: docs/builtins-reference.md
##########
@@ -127,12 +123,12 @@ kmeans(x, centers, iter.max, nstart)
### Arguments
-| Name | Type | Default | Description |
-| :---- | :------------- | -------- | :------------------------------- |
-| x | Matrix[Numeric] | required |The input Matrix to do KMeans on..|
-|centers | Int | -------- |The no. of cenetrs |
-|iter.max|Int |10 |Max no. of iterations allowed |
-|nstart |Int |10 |No. of random starting positions |
+| Name | Type | Default | Description |
+| :-- | :-- | -- | :-- |
Review comment:
Done. Only the changes to 'centers' were done this time.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [systemml] t201 commented on a change in pull request #933: [DOC] Documentation for builtin KMeans function
Posted by GitBox <gi...@apache.org>.
t201 commented on a change in pull request #933:
URL: https://github.com/apache/systemml/pull/933#discussion_r434601596
##########
File path: docs/builtins-reference.md
##########
@@ -143,12 +139,13 @@ kmeans(x, centers, iter.max, nstart)
| String |The output matrix with the centroids |
+
+**DML-bodied built-in functions** are written as DML-Scripts and executed as such when called.
### Example
-KMeans = rand(Matrix x, centers k=0,int iter.max=25,int nstart=10)
-return(Matrix[numerical] C,Matrix[numerical] Y)
+```h2o.kmeans(x = predictors, k = 100, estimate_k = T, standardize = F,
Review comment:
The syntax errors are quite persistent . No example worked in the collab till now.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [systemml] j143 commented on a change in pull request #933: [DOC] Documentation for builtin KMeans function
Posted by GitBox <gi...@apache.org>.
j143 commented on a change in pull request #933:
URL: https://github.com/apache/systemml/pull/933#discussion_r434760476
##########
File path: docs/builtins-reference.md
##########
@@ -143,12 +139,13 @@ kmeans(x, centers, iter.max, nstart)
| String |The output matrix with the centroids |
+
+**DML-bodied built-in functions** are written as DML-Scripts and executed as such when called.
### Example
-KMeans = rand(Matrix x, centers k=0,int iter.max=25,int nstart=10)
-return(Matrix[numerical] C,Matrix[numerical] Y)
+```h2o.kmeans(x = predictors, k = 100, estimate_k = T, standardize = F,
Review comment:
Kindly have a look at how other functions are documented.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [systemml] j143 commented on pull request #933: [DOC] Documentation for builtin KMeans function
Posted by GitBox <gi...@apache.org>.
j143 commented on pull request #933:
URL: https://github.com/apache/systemml/pull/933#issuecomment-638598329
Thank you, @t201 - LGTM. 👍
🎉 🚀
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org