Feature map(xgb.fmap) and XGBoost dump file(xgb.dump) upload request.

[SimonZhao777]

Hi Far0n, I am currently using XGBoost to select feature interaction for my LR model, I find your work very interesting, however I have encountered a problem to get the XgbFeatureInteractions.exe runing with my xgb.dump file. I am using scala language and XGBoost4j package running on Spark version 1.6, so I am not sure if the XGBoost dump file that my project created is having a different format from yours, so it would be really nice if you can upload the files xgb.fmap and xgb.dump that were mentioned in README.md in your project so that I can do a format check to see what wrong with my dump file. Thank you.

[SimonZhao777]

Problem solved, I have figured it out, turns out my xgb.dump lacks tree number lines. For someone who might be interested how I solved this dump file problem: Here is the code I wrote in Scala: ` def createFeatureMap(savePath:String, featureNames:Array[String]): Unit = {

val writer = new PrintWriter(savePath, "UTF-8") for (i <- 0 until featureNames.length) { writer.print(i + "\t" + featureNames(i) + "\t" + "q\n") } writer.close() }

val modelInfos = xgBoostModel.booster.getModelDump("../xgb.fmap", withStats = true)

def saveDumpModel(modelPath: String, modelInfos: Array[String]): Unit = { val writer = new PrintWriter(modelPath, "UTF-8") for (i <- 0 until modelInfos.length) { writer.print(s"booster[$i]:\n") writer.print(modelInfos(i)) } writer.close() } ` XGBDumpFile.zip

You can firstly call createFeatureMap and the feature map will be saved in specified location. Then prepare your XGBoost model dump(in the code above is modelInfos) and pass it into saveDumpModel Function then you will get a dump file that can be parsed by XGBFI.

The attached files are my created xgb.fmap and xgb.dump just to give you a reference.

[Far0n]

Thx for your feedback @SimonZhao777. I'll have a look at the dump and try to make xgbfi compatible with this format.

[SimonZhao777]

@Far0n The attached xgb.dump file is already compatible with your solution, the reason why my former version didn't work was because the code did not contain this line "writer.print(s"booster[$i]:\n")" in function saveDumpModel, so after adding this line, the dump file is working fine with xgbfi. Thank you for work!