{
  "id": 371528,
  "title": "Is there any way to identify which feature to and which feature to exclude without domain knowledge?",
  "url": "/competitions/g2net-detecting-continuous-gravitational-waves/discussion/371528",
  "author_name": "Mohit Sharma",
  "post_date": "2022-12-10T16:52:31.954000",
  "votes": 1,
  "comment_count": 1,
  "views": 0,
  "content": "<p>I have applied feature selection and Extraction techniques but still not able to interpret features based on business usecases. </p>",
  "messages": [
    {
      "id": 2061129,
      "postDate": "2022-12-10T17:52:59.107Z",
      "content": "<p>You can try RPart using the Caret package in R, but you need compute the features in a CSV first:</p>\n<pre><code>\ntrain_df  read_csv header row.names  \nsubsets   \nset.seed\n\nctrl.rfe.rpart  rfeControlfunctionscaretFuncs method   \n                             number   returnResamp   verbose  \n\nrf.rfe.rpart  rfeTarget. datatrain_dfsizessubsets\n                    rfeControlctrl.rfe.rpart method  \n\nrf.rfe.rpart\nrf.rfe.rpartfit\nrf.rfe.rpartoptsize\nrf.rfe.rpartoptVariables\n</code></pre>\n<p>I hope this helps :)</p>\n<p>Cheers</p>",
      "rawMarkdown": "You can try RPart using the Caret package in R, but you need compute the features in a CSV first:\n\n```r\n###Rpart example\ntrain_df <- read_csv(\"your_computed_features.csv\", header=TRUE, row.names = 1)\nsubsets <- c(1:12) #the columns subsets with the features to analyze\nset.seed(3456)\n\nctrl.rfe.rpart <- rfeControl(functions=caretFuncs, method = \"cv\", \n                             number = 5, returnResamp = \"final\", verbose = TRUE)\n\nrf.rfe.rpart <- rfe(Target~., data=train_df,sizes=subsets,\n                    rfeControl=ctrl.rfe.rpart, method = \"rpart\")\n\nrf.rfe.rpart\nrf.rfe.rpart$fit\nrf.rfe.rpart$optsize\nrf.rfe.rpart$optVariables\n```\n\nI hope this helps :)\n\nCheers",
      "votes": 1
    },
    {
      "id": 2061055,
      "postDate": "2022-12-10T16:52:31.953Z",
      "content": "<p>I have applied feature selection and Extraction techniques but still not able to interpret features based on business usecases. </p>",
      "rawMarkdown": "I have applied feature selection and Extraction techniques but still not able to interpret features based on business usecases. ",
      "votes": 1
    }
  ],
  "comments": [
    {
      "id": 2061129,
      "author_name": "Zollkron",
      "author_url": "",
      "post_date": "2022-12-10T17:52:59.107000",
      "content": "<p>You can try RPart using the Caret package in R, but you need compute the features in a CSV first:</p>\n<pre><code>\ntrain_df  read_csv header row.names  \nsubsets   \nset.seed\n\nctrl.rfe.rpart  rfeControlfunctionscaretFuncs method   \n                             number   returnResamp   verbose  \n\nrf.rfe.rpart  rfeTarget. datatrain_dfsizessubsets\n                    rfeControlctrl.rfe.rpart method  \n\nrf.rfe.rpart\nrf.rfe.rpartfit\nrf.rfe.rpartoptsize\nrf.rfe.rpartoptVariables\n</code></pre>\n<p>I hope this helps :)</p>\n<p>Cheers</p>",
      "votes": 1,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2061129": "You can try RPart using the Caret package in R, but you need compute the features in a CSV first:\n\n```r\n###Rpart example\ntrain_df <- read_csv(\"your_computed_features.csv\", header=TRUE, row.names = 1)\nsubsets <- c(1:12) #the columns subsets with the features to analyze\nset.seed(3456)\n\nctrl.rfe.rpart <- rfeControl(functions=caretFuncs, method = \"cv\", \n                             number = 5, returnResamp = \"final\", verbose = TRUE)\n\nrf.rfe.rpart <- rfe(Target~., data=train_df,sizes=subsets,\n                    rfeControl=ctrl.rfe.rpart, method = \"rpart\")\n\nrf.rfe.rpart\nrf.rfe.rpart$fit\nrf.rfe.rpart$optsize\nrf.rfe.rpart$optVariables\n```\n\nI hope this helps :)\n\nCheers",
    "2061055": "I have applied feature selection and Extraction techniques but still not able to interpret features based on business usecases. "
  }
}