Gradient boosting tutorial never actually touches X_train and y_train

[mobiusklein]

Pre-issue creation checklist

• No duplicates of this issue have already been created.

Name of the tutorial/dataset

NIST (part 2): Traditional ML: Gradient Boosting

Describe the mistake or typo

Gradient boosting tutorial never actually touches X_train and y_train

Additional context

Lines using X_test when context suggests it should be X_train

ProteomicsML/tutorials/fragmentation/_nist-2-traditional-ml-gradient-boosting.ipynb

Lines 402 to 412 in 512dc93

	"outputs": [],
	"source": [
	"reg = GradientBoostingRegressor()\n",
	"\n",
	"X_train = train_val_encoded.drop(columns=[\"spectrum_id\", \"b_target\", \"y_target\"])\n",
	"y_train = train_val_encoded[\"b_target\"]\n",
	"X_test = test_encoded.drop(columns=[\"spectrum_id\", \"b_target\", \"y_target\"])\n",
	"y_test = test_encoded[\"b_target\"]\n",
	"\n",
	"reg.fit(X_test, y_test)"
	]

ProteomicsML/tutorials/fragmentation/_nist-2-traditional-ml-gradient-boosting.ipynb

Lines 442 to 452 in 512dc93

	"cell_type": "code",
	"execution_count": null,
	"metadata": {},
	"outputs": [],
	"source": [
	"def objective(n_estimators):\n",
	" # Define algorithm\n",
	" reg = GradientBoostingRegressor(n_estimators=n_estimators)\n",
	"\n",
	" # Fit model\n",
	" reg.fit(X_test, y_test)\n",

Also, this may be a failure in my reading/understanding the notebook, but the notebook says the model it trains can only predict y ions, but the training target is listed as b_target, not y_target (see line 407) in first snippet.

ProteomicsML/tutorials/fragmentation/_nist-2-traditional-ml-gradient-boosting.ipynb

Line 619 in 512dc93

"And of course, this model can only predict y-ion intensities. You can repeat the\n",