omarfsosa

"""
Binomial Asset Pricing Model
----------------------------

These are some class implementations I'm using to replicate in code
some of the examples presented in Shreve's book on Stochastic Calculus
part 1.


Example usage (example 1.3.1)

omarfsosa

from tqdm.auto import tqdm

# The example below is for a classification model
# with the last layer of the model producing the logits
# --i.e. without a softmax layer 

def train_sinle_epoch(dataloader, model, loss_fn, optimizer):
    correct_total = 0
    size_total = 0
    with tqdm(dataloader, unit="batch") as tepoch:

omarfsosa

#!/bin/bash
pip install matplotlib numpy pandas scipy scikit-learn
pip install jax numpyro

omarfsosa

"""
Approximate inverse CDF for the poisson distribution in Jax.
Based on the approximation method proposed in [1].

References
----------
[1]: https://people.maths.ox.ac.uk/gilesm/codes/poissinv/paper.pdf
"""
import jax
import jax.numpy as jnp

omarfsosa

python -m pip install --upgrade pip
pip install cython pybind11 pythran numpy
OPENBLAS=$(brew --prefix openblas) CFLAGS="-falign-functions=8 ${CFLAGS}" pip install --no-use-pep517 scipy==1.7.0
pip install pandas matplotlib jupyter jupyterlab
OPENBLAS=$(brew --prefix openblas) CFLAGS="-falign-functions=8 ${CFLAGS}" pip install --no-use-pep517 scikit-learn==1.0.1
jupyter nbextension enable --py widgetsnbextension

omarfsosa

"""Visualization tools"""
import matplotlib.pyplot as plt
import numpy as np


def spaghetti_plot(
    x, y, n_samples=20, indices=None, ax=None, plot_kwargs=None
):
    """
    Plots x against a few picked examples of y.

omarfsosa

4.69
5.85
7.28
3.49
3.26
5.69
12.13
10.42
5.28
5.88

omarfsosa

functions {
    real normalisation(real lam, int lb, int ub) {
        real result = 0;
        for (x in lb:ub) {
            result += (lam^x)/tgamma(x+1);
        }
        result *= exp(-lam);
        return result;
    }
}

omarfsosa

conda create -n myenv python=3.6
conda activate myenv
pip install ipykernel
python -m ipykernel install --user --name myenv --display-name "Python 3.6 (myenv)"

omarfsosa

import numpy as np

def neg_binomial_2_rng(mu, phi, size=None):
    """
    RNG for the negative binomial parametrised in terms
    of the mean and the precision.
    mean = mu
    variance = mu + (mu**2)/phi
    """
    tau = np.random.gamma(phi, mu/phi, size=size)