fabiommaia Goto Github PK

ann_from_scratch

Implementation of a Neural Network from scratch with a customized pipeline to parameter tunning and cross validation

anomalies_categorics

autotune_cnn_transferlearning_adaptation

bayesian_inference

celebnet

Training CNN models using CelebA with two approachs: AlexNet with regular train and MobileNetv2 with transfer learning

classification---celeba

Implementation of 4 models types: Multinomial Logistic Regression, Suport Vector Machines, Multilayer Perceptron and Convolutional Neural Networks considering classification task of human faces

face_recognition_4models

Face recognition task using classical approaches implemented from scratch. Later comparison with CNN and more robust techniques.

finance_ml

fjsp_genetic_tabu

Application of hybrid genetic algorithm and tabu search for flexible job shop scheduling problem (FJSP)

genetic_method_supervised_learning

Comparison between genetic algorithm (using scipy lib) versus Logistic Regression on cancer dataset to find best parameters

kedro_tutorial

llms

machine_learning

Simple implementation of most common machine learning algorithms

model-analyser

Functions related to model understanding, such as feature importance description and general data analysis tools

ner-portuguese-br

reading_tables_documents

Detecting tables, shapes, and splitting tables into columns for further OCR using OpenCV

sanity_checker

Implementation of general functions to understand data drift between samples with practical use

svm_from_scratch

SVM implementation from scratch, using quadratic programming to solve dual problem

template_match

Subtracting background of documents using opencv and template match

trading-predictions

Use of classical machine learning methods to predict fluctuations in trading using yahoo finance as data source

word2vec_visualization