/AI-python-connect

Primary LanguageJupyter Notebook

머신러닝/딥러닝을 위한 Python

강의개요

본 강의는 머신러닝, 딥러닝을 배우기 위해 기본적으로 이해해야하는 Python을 다룹니다. 빠른 시간 내에 Python 기초 문법을 복습하고 머신러닝, 딥러닝의 근간을 이루는 Numpy, Pandas와 친숙해지고 싶은 분에게 추천합니다. 참고 - 머신러닝/딥러닝을 위한 Python

강의정보

강의구성

Chapter 0 - Environment setup

Package installation
conda create -n ml_python python=3.5
conda install numpy seaborn scikit-learn jupyter
conda install nltk gensim matplotlib

Chapter 1 - Pythonic Code

Chapter 2 - Numpy section

  • Numpy overview
  • ndarray
  • Handling shape
  • Indexing & Slicing
  • Creation functions
  • Opertaion functions
  • Array operations
  • Comparisons
  • Boolean & fancy Index
  • Numpy data i/o
  • Assignment: Numpy in a nutshell

Chapter 3 - Pandas section

  • Pandas overview
  • Series
  • DataFrame
  • Selection & Drop
  • Dataframe operations
  • lambda, map apply
  • Pandas builit-in functions
  • Lab Assignment: Build a matrix
  • Groupby I
  • Groupby II
  • Casestudy
  • Pivot table & Crosstab
  • Merg & Concat
  • Database connection & Persistance

Chapter 4 - OOP section

  • Objective oriented programming overview
  • Objects in Python
  • Lab: Note and Notebook
  • OOP characteristics
  • Decorators, Static And Class Methods
  • Abstract Classes

Chapter 5 - Linear regression

  • Linear regression overview
  • Cost functions
  • Linear Equality
  • Gradient descent approach
  • Linear regression wtih gradient descent
  • Linear regression wtih Numpy
  • Multivariate linear regression models
  • Multivariate linear regression with NumPy
    • Regularization - L1 and L2
  • Implementation of generalization with NumPy
  • Linear regression with sklearn

Chapter 6 - Logistic regression

  • Logistic regression overview
  • Sigmoid function
  • Cost function
  • Logistic regression implementation with Numpy
  • Maximum Likelihood estimation
  • Regularization problems
  • Logistic regresion with sklearn
  • Softmax fucntion for Multi-class classification
  • Cross entropy loss function
  • Softmax Logistic Regression
  • Performance measures for classification

References