Skip to content
View mehdi-irema's full-sized avatar
💭
Learning new stuff!
💭
Learning new stuff!
  • USI Università della Svizzera italiana
  • Lugano, Switzerland

Block or report mehdi-irema

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
mehdi-irema/README.md

Hi, I'm M. Mahdi Ameridaragheh

I'm an MSc student in Statistical Sciences (Data Science) at the University of Bologna, currently at USI Lugano on the UNIBO–USI double degree. I like problems where careful statistics meets modern machine learning, especially interpretable ML for health and wearable-sensor data.

Currently

  • MSc thesis (USI): multimodal Concept Bottleneck Models with continuous concepts, recognising affect and cognitive load from wrist-worn and smart-glasses signals.

Selected projects

Project What Stack
breast-cancer-metastasis-prediction BSc thesis: predicts metastasis or recurrence from 3,753 clinical records (Motamed Cancer Institute) with seven ML models, using a leakage-safe pipeline with SMOTE-ENN. Data not shared. Python · scikit-learn · imbalanced-learn
garment-packshot-pipeline Neural Wave 2026 hackathon (team). Turns amateur phone photos of garments into studio packshots: multi-view generative editing, then picks the best candidate by colour consensus in Lab space and mask symmetry and tilt. Python · NumPy · scikit-image
spam-detection-R Ten supervised learners on the Spambase data, all tuned by cross-validation. Boosting reaches about 5% test error. R · gbm · randomForest · glmnet · e1071
uranium-score-driven-model A robust Student-t score-driven (DCS-t) model vs ARIMA for the global uranium price. Multi-start MLE finds a 100-point AIC improvement. R · own MLE implementation
les-miserables-network-analysis Network analysis of a character co-appearance graph: centrality, clustering, community detection, assortativity. R · igraph
deep-learning-specialization From NumPy networks built from scratch to CNNs, sequence models and Transformers. Python · TensorFlow/Keras

Background

  • BSc in Industrial Engineering (ranked 5th of 83)
  • BSc thesis: machine learning for breast-cancer metastasis prediction (supervised by Dr. Farnaz Barzinpour, IUST)
  • Interested in player-behaviour analytics and recommender systems

Tools

Python (PyTorch, TensorFlow/Keras, scikit-learn, pandas, NumPy) · R (igraph, glmnet, randomForest) · SAS · Tableau · Jupyter · Git

Contact

LinkedIn · mahdi.ameri130@gmail.com

Pinned Loading

  1. breast-cancer-metastasis-prediction breast-cancer-metastasis-prediction Public

    BSc thesis: predicting breast-cancer metastasis/recurrence from 3,753 clinical records with 7 ML models; leakage-safe imblearn pipeline (data not included)

    Python

  2. garment-packshot-pipeline garment-packshot-pipeline Public

    Neural Wave 2026 hackathon: phone photos of garments to studio packshots via generative editing + colour/shape consensus selection

    Python

  3. spam-detection-R spam-detection-R Public

    Comparing 10 supervised learning methods in R (boosting, random forests, SVM, LASSO, KNN, trees...) on the UCI Spambase data, tuned by cross-validation

    R

  4. uranium-score-driven-model uranium-score-driven-model Public

    Robust Student-t score-driven (DCS-t) model vs ARIMA for the global uranium price (FRED PURANUSDM, 1992-2026), in R

    R

  5. les-miserables-network-analysis les-miserables-network-analysis Public

    Descriptive network analysis of the Les Misérables co-appearance network in R/igraph: centrality, clustering, communities, assortativity

  6. deep-learning-specialization deep-learning-specialization Public

    My work for the Deep Learning Specialization (DeepLearning.AI): neural networks, CNNs, sequence models, tuning and regularisation.

    Jupyter Notebook 1