# My NYU Course Materials

teardown

machine-learning

My lecture notes & resources from Machine Learning in Finance. It is more likely you will drown rather than starve when it comes to data science and ML resources. Nonetheless, access to all my lecture notes and notebooks are below. Email me if you would like permission to access the materials on Google Drive.

Module | Sub-Module | Topic | Description | Material Type | Filename |

1-Link | 1a | ML in Finance | Course movtivation:What is the state of Machine Learning and why is it important. What are some use cases in Finance and how is it used in practice in the workplace | PPT | Intro to ML in Finance |

FinTech Landscape | Primer on the FinTech | PPT | FinTech Landscape | ||

1b | Introduction to Python | Review of notebooks from [JVP] Chapters 1-3; also used [WM] | Jupyter Notebook | 0_intro_to_python | |

First sklearn model | Based on [JVP], framework for instantiating, fitting, and predicting a model in Python | Jupyter Notebook | 1_first_model | ||

Model Selection | Description of hypterparameter selection using cross-validation; bias-variance trade-off. | PPT | Model Selection | ||

Cross-validation illustrated | Builds from: all in-sample, train-test split, cross-validation, gridsearch cross validation | Jupyter Notebook | 2_crossval | ||

Bias-Variance | Demonstrates concept of overfitting (low-bias, high variance) and the conceptual balance to found ideally through gridsearch cross validation | Jupyter Notebook | 3_bias_variance | ||

2-Link | 2a | Linear Regression | A review of linear regression derivation motivating regularized regression. | Linear Regression | |

2b | Review of Value-at-Risk | A review of Value-at-Risk to help the regularized regression use case. | PPT | Value-at-Risk | |

VaR hedging | Demonstration of how one would find securities to hedge a portfolio VaR | Jupyter Notebook | VaR Hedging | ||

2c | Basket Replication | Demonstration of how one would find securities to replicate an instrument. | Jupyter Notebook | Basket Replication | |

2-data | Portfolio Prices | Unknown time series (was a single name stock) representing a portfolio PnL. | Data | pnl_prices.csv | |

2-data | Hedge Prices | Single name stock prices available for shorting. | Data | hedge_prices.csv | |

3-Link | 3a | Decision Trees & Random Forests: Ensembles | An overview from difference resources on Decision Trees and Random Forest | Ensembles.pdf | |

3b | Decision Trees Notebook | Adaptation from A Geron's notebook that accompanies his textbook | Jupyter Notebook | decision_trees | |

3c | Ensemble & Random Forest Notebook | Adaptation from A Geron's notebook that accompanies his textbook | Jupyter Notebook | ensembles_random_forest | |

3d | Trading Model | Framework for creating a Machine Learning trading strategy. | Jupyter Notebook | ML_trading_model | |

4-Link | 4a | Methods in Unsupervised Learning | An overview of unsupervised learning, k-means and hierarchical clustering. | unsupervised_learning.pdf | |

4b | FX Use Case | Demonstration of clustering currencies by their average and std dev of returns. | Jupyter Notebook | kmeans_FXdata.ipynb | |

4c | Agglomerative Clustering | Demonstration of agglomerative clustering | Jupyter Notebook | hierarchical_clustering.ipynb | |

4d | Hierarchical Risk Parity | Implementation of Hierarchical Risk Parity. | Jupyter Notebook | hierarchical_risk_parity.ipynb |