sankalp-indish/alexithymia-adhd-predictor

🧠 Alexithymia & ADHD Predictor

A probabilistic MLP+SMOTED SVM Based classifier for early screening insights into emotional awareness and attention-related traits.

🧩 Model Details

• Model Type: bertopic-based transformer classifier
• Dataset: Alexithymia-ADHD
• Task: text-classification
• Languages: English 🇬🇧
• Metrics Tracked: Accuracy, Precision, Recall, F1

---

🔍 Use Cases

• 🧠 Early research-based screening tool
• 📊 Self-reflection on attention/emotion patterns
• 🧪 Experimental psychology & behavioral NLP

---

🚀 Quick Start

from transformers import pipeline

classifier = pipeline("text-classification", model="sankalp-indish/alexithymia-adhd-predictor")
classifier("I often find it hard to describe how I feel.")

📈 Example Output

The model takes in short user inputs and returns a probabilistic score for both Alexithymia and ADHD traits.

Input Text	Alexithymia Score	ADHD Score
"I often struggle to explain what I'm feeling."	1	0
"I get distracted easily and can't finish tasks on time."	0	1
"Sometimes I can't tell if I'm anxious or excited."	1	0
"I daydream a lot and forget what I'm doing."	0	1

🧪 These values are probabilities, not diagnoses. Scores above 0.5 and 0.3 (based on the availability of the dataset) suggest strong presence of that trait.

🧠 Behind the Model

1. This classifier is trained on combined NLP embeddings from real-world psychological scale responses. It uses probabilistic classification to predict traits, not diagnoses. And currently is scale and input of text based
2. Framework: bertopic + sklearn
3. Fine-tuned over custom-labeled dataset
4. Designed for interpretability & simplicity

⚠️ Disclaimer

This tool is not a substitute for clinical diagnosis. It is provided for research and educational use only. Please consult a qualified mental health professional for clinical assessment.

👨‍💻 Author

Developed by Sankalp Indish [sankalp_indish] https://huggingface.co/sankalp-indish