Upload README.md with huggingface_hub

README.md

@@ -1,50 +1,5 @@
----
-library_name: scikit-learn
-license: other
-language: en
-tags: [tabular, classification, random-forest, tourism]
-datasets: [deva8217/tourism-wellness]
----
-
 # Wellness Tourism Purchase Model
 
-**Best model:** RandomForest  
-**Validation F1:** 0.7429
-
-**Test metrics**
-- ROC-AUC: 0.9482  
-- AUPRC: 0.8454  
-- Accuracy: 0.9068  
-- Precision: 0.7440  
-- Recall: 0.7862  
-- F1: 0.7646  
-**Threshold:** 0.32
-
-## How to use
-```python
-from huggingface_hub import hf_hub_download
-import joblib, pandas as pd
-
-# download model artifact
-p = hf_hub_download("deva8217/tourism-wellness-model", "best_model.joblib")
-model = joblib.load(p)
-
-# df must contain the same training columns (names & dtypes)
-df = pd.DataFrame([{
-    "Age": 44, "CityTier": 1, "Passport": 1, "OwnCar": 1,
-    "TypeofContact": "Company Invited", "Occupation": "Salaried",
-    "Gender": "Male", "NumberOfPersonVisiting": 4, "PreferredPropertyStar": 5,
-    "MaritalStatus": "Married", "NumberOfTrips": 2, "NumberOfFollowups": 2,
-    "Designation": "Executive", "MonthlyIncome": 50000,
-    "PitchSatisfactionScore": 7, "ProductPitched": "Basic",
-    "DurationOfPitch": 30, "NumberOfChildrenVisiting": 0
-}])
-
-proba = model.predict_proba(df)[:, 1][0]
-print("purchase_probability:", round(float(proba), 4))
-
-# Wellness Tourism Purchase Model
 Best model: **RandomForest**  
-Validation F1: **0.7429**  
-Test metrics: {"roc_auc": 0.9482475743260445, "auprc": 0.8453597025735655, "accuracy": 0.9067796610169492, "precision": 0.7440476190476191, "recall": 0.7861635220125787, "f1": 0.764525993883792}  
-Threshold: **0.32**
+Test metrics: {"roc_auc": 0.9390964894911036, "auprc": 0.8021290736713121, "accuracy": 0.8995157384987893, "precision": 0.7065217391304348, "recall": 0.8176100628930818, "f1": 0.7580174927113703, "threshold": 0.3497817045767868}  
+Threshold: **0.35**