Maji-xgb-model

🌀 Maji Archetype Regressor

This model is part of the Soulprint Archetype Framework, which measures cultural and emotional dimensions of communication.

The Maji archetype represents healing, restoration, and emotional grounding.

• Model type: XGBoost regressor

• Input: English text (1–5 sentences)

• Output: A continuous score from 0.0–1.0

  • 0.0–0.2 → Disconnected, dismissive, tense

  • 0.2–0.4 → Mild support, low grounding

  • 0.4–0.6 → Neutral, steadying presence

  • 0.6–0.8 → Warm, supportive, moderately healing

  • 0.8–1.0 → Deeply restorative, profound comfort

📊 Training

• Dataset size: ~1,070 rows

• Balanced bins: Each 0.2 range has ~210 rows

• Embedding model: sentence-transformers/all-mpnet-base-v2

• Regressor: XGBoost

• Best config:

  • n_estimators=500

  • learning_rate=0.04

  • max_depth=4

  • subsample=0.8

  • colsample_bytree=0.8

🧪 Results

• MSE: 0.0144

• R²: 0.803

🚀 Usage

import xgboost as xgb
from sentence_transformers import SentenceTransformer
from huggingface_hub import hf_hub_download

# -----------------------------
# 1. Download model from Hugging Face Hub
# -----------------------------
REPO_ID = "mjpsm/Maji-xgb-model"  # replace if you use a different repo name
FILENAME = "Maji_xgb_model.json"

model_path = hf_hub_download(repo_id=REPO_ID, filename=FILENAME)

# -----------------------------
# 2. Load model + embedder
# -----------------------------
model = xgb.XGBRegressor()
model.load_model(model_path)

embedder = SentenceTransformer("all-mpnet-base-v2")

# -----------------------------
# 3. Example prediction
# -----------------------------
text = "I held my sister’s hand until she calmed down."
embedding = embedder.encode([text])
score = model.predict(embedding)[0]

print("Predicted Maji Score:", round(float(score), 3))

🌍 Applications

• Measuring healing energy in dialogue

• Analyzing cultural communication through Soulprint

• Integrating into chatbots, dashboards, or community storytelling tools