I am trying to use a HDBSCAN model that I previously pkl'ed from google co-lab over to a local python project I am doing. The problem I am having is that no matter what this model will not generate any prediction data. I believe that I have trained the model correctly and the example array is what the input data looks like. Any help is greatly appreciated. TESTING THE PREDICTIONS import joblib import numpy as np import hdbscan model = joblib.load("hdbscan_model.pkl") print("Model loaded") test_sample = [ 1, 516, 1, 237, 1, -667, 0, 945, 1, -2882, 1, 1519, 1, -2878] try: labels, strengths = hdbscan.approximate_predict(model, test_sample) print("Predicted labels:", labels) except Exception as e: print("Prediction error:", e) HOW MODEL WAS TRAINED IN COLAB X = df_cleaned.drop(columns=['Failure_area']) #Scale features X_scaled = StandardScaler().fit_transform(X) #UMAP for 2D projection reducer = umap.UMAP(random_state=42) X_umap = reducer.fit_transform(X_scaled) # HDBSCAN clustering clusterer = hdbscan.HDBSCAN(min_cluster_size=40, min_samples= 55, prediction_data=True, gen_min_span_tree=True) import pickle # Save the model to a file with open('hdbscan_model.pkl', 'wb') as f: pickle.dump(clusterer, f) Continue reading...
