Report various statistics stemming from a confusion matrix in a tidy fashion. 🎯

specificity tidy sensitivity kappa confusion-matrix rand fallout false-discovery-rate f1-score dice-coefficient precision-recall positive-predictive-value negative-predictive-value true-positive-rate true-negative-rate false-positive-rate false-negative-rate balanced-accuracy false-omission-rate peirce

• Updated Jul 10, 2020
• R

Emotion Analysis with Transformers

sentiment-analysis metrics tokenizer pytorch p-value emotion-analysis nlp-machine-learning word2vec-embeddinngs balanced-accuracy distilbert huggingface-transformers friedman-test matthews-correlation-coefficient roberta-large custom-bert-model dataloader-constructor models-ranking

• Updated Feb 24, 2026
• Jupyter Notebook

This repository contains the computational project developed for the Data Analysis Statistics (EAD) course. The primary objective is to explore, segment, and classify higher education institutions as Public or Private (bl_private) based on academic, demographic, and financial indicators.

cross-validation pca-analysis supervised-learning logistic-regression preprocessing feature-engineering predictive-modeling lda unsupervised-learning kmeans-clustering f1-score knn-classification elbow-method qda auc-roc-score balanced-accuracy silhouette-method

• Updated Jun 3, 2026
• R

Multimodal BBQ (bias) visual question answering: pick the right answer from image + context + question + 3 options, and answer "unknown" when evidence is insufficient. Offline, open-weights Qwen2.5-VL (4-bit) + balanced chain-of-thought. Public Balanced Accuracy 0.883.

pytorch kaggle vqa bias bbq vlm visual-question-answering multimodal vision-language dacon balanced-accuracy huggingface-transformers uncertainty-calibration llm prompt-engineering chain-of-thought offline-inference qwen2-5-vl 4-bit-quantization open-weights

• Updated Jun 22, 2026
• Python