msellamitn/data-science-tutorials

data-science-tutorials

This repository contains jupyter-notebooks to accompany the tutorials for our data science lectures. The following topics are covered (each within a separate folder).

1. Dataset Visualization (Boston Housing minus the linear regression; also other datasets like Flower, MNIST-digits, 20newsgroups) working/visualizing one dataset (incl. Matplotlib; .describe attribute; box-plot, min-max-normilization; boston housing; linear reg c/o dsP)
2. Clustering
3. Association Rule Learning (dataset yet to be determined; preferably from scikit learn)
4. Regression (linear regression from Boston Housing and Car Prices)
5. Bayes Learning (for spam filtering/text classification)
6. Classification with Decision Trees (start with small 5-line dataset)
7. Neural Networks (use keras.io to build a neural network for MNIST-digit classification) keras (for MNIST class); OPT gensim (for word2vec; pick dataset from tensorflow); then auto-encoder for representatino learning
8. OPTIONAL MapReduce

Packages

See our python-tutorials on instructions how to set this up on your machine.

required

• Python (>= 2.7 or >= 3.3)
• NumPy (>= 1.6.1)
• SciPy (>= 0.9)
• scikit-learn (>=0.18.1); documentation, also as pdf with Quick Start and Tutorials
• Matplotlib >= 2.1.1

optional

• Pandas; [documentation] also as pdf

Table of contents

• 0-Intro

  • Scikit-learn-overview.ipynb
  • Web Mining Project .ipynb

• 1-Datasets_Visualization_and_preprocessing

  • 1-IRIS.ipynb
  • 2-Boston_house_dataset.ipynb
  • 3-MNIST.ipynb
  • 4-UCI_CAR.ipynb
  • 5-20newsgroups.ipynb
  • 6-KDD_cup_2000_data_set.ipynb
  • Crawling_twitter_with_python.ipynb
  • MDS_projection.ipynb (IRIS)
  • PCA_projection.ipynb (IRIS)
  • scikit-learn-overview-and-preprocessing.ipynb (IRIS)
  • VA-InformationVisualisation-with-JavaScript-and-3DJs.ipynb
  • TODO try visualization with Orange (available through the conda-forge channel)

• 2-Clustering

  • Clustering_overview.ipynb (IRIS) (MNIST)
  • Tutorial_clustering_for_outlier_detection_3D.ipynb (Kddcup 1999)
  • Tutorial_clustering_for_outlier_detection.ipynb (Kddcup 1999)

• 3-Association-Rules

  • Apriori_asaini.ipynb (MBE_dataset)
  • Apriori.ipynb (Boston house)
  • Apriori_server.ipynb (Mango_dataset)
  • Assignment_Association_rule_learning.ipynb
  • Tutorial_association_rule_learning_shopping_basket.ipynb (KDDcup 2000)

• 4-Linear_regression_and_logistic_regression

  • Assignment_Linear_Regression.ipynb
  • Assignment_Logistic_regression.ipynb (UCI_car)
  • Boston_house_Linear_Regression.ipynb (Boston house)
  • Linear_regression_diabetes_dataset.ipynb
  • Linear-Regression.ipynb (Boston house)
  • Logistic_regression.ipynb (IRIS)
  • Small_scale_linear_regression.ipynb (KDDcup)
  • Supervised_Learning_with_Linear_Models.ipynb (Boston house)

• 5-KNN_classification

  • KNN_classification.ipynb (IRIS)
  • Metrics.ipynb (IRIS)

• 6-Bayes-Learning

  • Bayes-Learning.ipynb (IRIS) (20 news group)

• 7-Decision-Trees.ipynb (UCI_car)

• 8-Neural-Networks

  • keras-mnist.ipynb (MNIST)
  • Simple-NN.ipynb (make_moons)
  • Stacked-Denoising-Autoencoders.ipynb
  • INFO Software Comparison
    • keras.io (high-level, running on top of TensorFlow (default) or Theano) c/o Francois Chollet (written in Python)
    • Theano c/o Universite de Montreal (written in Python; tightly integrated with NumPy)
    • TensorFlow c/o Google Brain (written in Python/C++)

• 9-SVM

  • Assignment_SVM_for_OCR.ipynb (MNIST)
  • Support_Vector_Machines.ipynb (IRIS)

• A-Advanced_modules

  • NLP-with-NLTK-Short-Intro.ipynb

• B-Scripts

Links

Cheat Sheets

• cheat sheets for machine learning
• Cheat Sheet of Machine Learning and Python (and Math) Cheat Sheets
• Essential Cheat Sheets for Machine Learning and Deep Learning Engineers

Other Collections

• large repository of Data Science Python Notebooks

Module Specific

(should be listed at the module)

• Tutorials for TensorFlow