nicoloceneda / supervised-learning-models

Various supervised machine learning models

Supervised Learning Models

Author: Nicolo Ceneda
Contact: [email protected]
Website: nicoloceneda.github.io
Institution: Imperial College London
Course: PhD in Finance
Last update: 11 August 2020

   

Content of the Scripts:

│
├── 01_perceptron.py                          <-- Implementation of a single layer perceptron for bin-
│                                                 ary classification.
│
├── 01_perceptron_sl.py                       <-- Implementation of a single layer perceptron for mul-
│                                                 ti-class classification via scikit-learn.
│
├── 02_adaline_gd.py                          <-- Implementation of a single layer adaptive linear ne-
│                                                 uron for binary classification, via gradient descent 
│                                                 algorithm, with standardized features.
│
├── 02_adaline_sgd.py                         <-- Implementation of a single layer adaptive linear ne-
│                                                 uron for binary classification, via stochastic grad-
│                                                 ient descent algorithm, with standardized features.      
│
├── 03_logistic_regression_gd.py              <-- Implementation of a single layer logistic regression
│                                                 for binary classification, via gradient descent alg-
│                                                 orithm, with standardized features.                      
│
├── 03_logistic_regression_gd_sl.py           <-- Implementation of a single layer logistic regression 
│                                                 for multi-class classification, via gradient descent 
│                                                 algorithm, with standardized features, using scikit-
│                                                 learn.
│
├── 04_support_vector_machine_gd_sl.py        <-- Implementation of a support vector machine via scik- 
│                                                 it learn. 
│                                                
├── 05_decision_tree_sl.py                    <-- Implementation of a decision tree for multi-class c-
│                                                 lassification, with standardized features and gini
│                                                 impurity, using scikit-learn.
│                                              
├── 06_k_nearest_neighbors_sl.py              <-- Implementation of a k-nearest neighbors for multi-c-
│                                                 lass classification, with standardized features and 
│                                                 euclidean distance metric, using scikit-learn.
│                                                 
├── 07_mnist_dataset.py                       <-- Download mnist dataset and save the standardized fe-
│                                                 atures and class labels.
│                                                 
├── 08_multilayer_perceptron_gd.py            <-- Implementation of a multilayer perceptron for multi-
│                                                 class classification, with one hidden layer.
│                                                 
├── 08_multilayer_perceptron_gd_tf.py         <-- Implementation of a multilayer perceptron for multi-
│                                                 class classification, with two hidden layers, using 
│                                                 tensorflow.
│                                                 
├── 09_imdb_dataset.py                        <-- Download the imdb dataset and save the features and 
│                                                 targets.
│                                                 
├── 10_recurrent_neural_network_bi_lstm_tf.py <-- Implementation of a bidirectional lstm multilayer r-
│                                                 ecurrent neural network for sentiment analysis, with 
│                                                 a many-to-one architecture and two hidden layers, u-
│                                                 sing tensorflow.
│                                                
├── 10_recurrent_neural_network_simp_tf.py    <-- Implementation of a simple single layer recurrent n-
│                                                 eural network for sentiment analysis, with a many-to
│                                                 -one architecture and two hidden layers, using tens-
│                                                 orflow.
│                                                
├── 10_recurrent_neural_network_lstm_tf.py    <-- Implementation of a lstm multilayer recurrent neural 
│                                                 network for text generation, with a many-to-mani ar-
│                                                 chitecture and two hidden layers, using tensorflow.
│
└── images                                    <-- Images produced by the programs.