Thomas Wiecki twiecki

## hrp.py
import scipy.cluster.hierarchy as sch
import random
import numpy as np
import pandas as pd

def getIVP(cov, **kargs):
    # Compute the inverse-variance portfolio
    ivp = 1. / np.diag(cov)
    ivp /= ivp.sum()
    return ivp

## custom_rolling.py
import pandas as pd

from pandas.tseries.offsets import *

def rolling_window(data, resample='1BM', lookback=12 * BMonthEnd()):
    dts = data.resample(resample).index
    if lookback == 'aggregate':
        for dt in dts:
            yield data.loc[:dt]
    else:

## example_pymc3_model_cache.py
class BayesianModel(object):
    samples = 2000
    def __init__(self, cache_model=True):
        self.cached_model = None
        self.cached_start = None
        self.cached_sampler = None
        self.shared_vars = {}

    def cache_model(self, **inputs):
        self.shared_vars = self._create_shared_vars(**inputs)

## gist:f3494d4acf44bca79a31e62b5bd81807
def initialize(context):
    fetch_csv('URL',
              date_format='%d/%m/%y',
              universe_func=my_universe)

def my_universe(context, fetcher_data):
    my_stocks = set(fetcher_data['sid'])

    # log the size of the universe for debugging
    context.count = len(my_stocks)

## compute_cumulative_portfolio_returns_min.ipynb

      
              1 file
            
          
              0 forks
            
          
              0 comments
            
          
              0 stars
            
          
                twiecki
                / compute_cumulative_portfolio_returns_min.ipynb
            
            
              Created
              March 2, 2016 17:53
            
          
      Sorry, something went wrong. Reload?
      Sorry, we cannot display this file.
      Sorry, this file is invalid so it cannot be displayed.
      
          Viewer requires iframe.
      
    
## bayesian_neural_network.ipynb

      
              1 file
            
          
              7 forks
            
          
              9 comments
            
          
              27 stars
            
          
                twiecki
                / bayesian_neural_network.ipynb
            
            
              Last active
              February 22, 2022 01:28
            
              
                Bayesian Neural Network in PyMC3
              
          
      Sorry, something went wrong. Reload?
      Sorry, we cannot display this file.
      Sorry, this file is invalid so it cannot be displayed.
      
          Viewer requires iframe.
      
    
## abstract
All that glitters is not gold: Comparing backtest and out-of-sample performance of 800.000 trading algorithms

“Past performance is no guarantee of future returns”. This cautionary message will certainly match the experience of many
investors. When automated trading strategies are developed and evaluated using backtests on historical pricing data, there
is always a tendency, intentional or not, to overfit to the past. As a result, strategies that show fantastic performance on
historical data often flounder when deployed with real capital.

Quantopian is an online platform that allows users to develop, backtest, and trade algorithmic investing strategies. By
pooling all strategies developed on our platform we constructed a huge and unique data set of over 800.000 trading algorith
ms. Although we do not have access to source code, we have returns and portfolio allocations as well as the time the

## table.html
          <html>
          <table id="comparetable" class="clean">
           <tr>
            <td class="blank"> </td>
                                      <th>Machine Learning</th>
                                        <th>Frequentist<br>Statistics</th>
                                        <th>Probabilistic<br>Programming</th>
                                     </tr>
                                    <tr>
                <td class="rowTitle">Prediction</td>

## ppc.ipynb

      
              1 file
            
          
              0 forks
            
          
              3 comments
            
          
              0 stars
            
          
                twiecki
                / ppc.ipynb
            
            
              Last active
              August 29, 2015 14:25
            
              
                Example NB to run a PPC test using new random() methods.
              
          
      Sorry, something went wrong. Reload?
      Sorry, we cannot display this file.
      Sorry, this file is invalid so it cannot be displayed.
      
          Viewer requires iframe.
      
    
## exmachina.py
# After seeing Ex Machina I tweeted:
# "So disappointed in #ExMachina -- the #Python wasn't PEP8 compatible..."
# https://twitter.com/twiecki/status/599615426922938368
# For the original code in the movie as well as what it does (it's kinda neat), see:
# http://moviecode.tumblr.com/post/119171520870/in-the-movie-ex-machina-which-is-really-great

# To put my code where my mouth is, I decided to make the code PEP8 compatible which you can see below.
# I also did some minor cosmetic changes like changing i = i + 1 to i += 1
# Other ideas for improvements? Comment below.
	import scipy.cluster.hierarchy as sch
	import random
	import numpy as np
	import pandas as pd

	def getIVP(cov, **kargs):
	# Compute the inverse-variance portfolio
	ivp = 1. / np.diag(cov)
	ivp /= ivp.sum()
	return ivp
	import pandas as pd

	from pandas.tseries.offsets import *

	def rolling_window(data, resample='1BM', lookback=12 * BMonthEnd()):
	dts = data.resample(resample).index
	if lookback == 'aggregate':
	for dt in dts:
	yield data.loc[:dt]
	else:
	class BayesianModel(object):
	samples = 2000
	def __init__(self, cache_model=True):
	self.cached_model = None
	self.cached_start = None
	self.cached_sampler = None
	self.shared_vars = {}

	def cache_model(self, **inputs):
	self.shared_vars = self._create_shared_vars(**inputs)
	def initialize(context):
	fetch_csv('URL',
	date_format='%d/%m/%y',
	universe_func=my_universe)

	def my_universe(context, fetcher_data):
	my_stocks = set(fetcher_data['sid'])

	# log the size of the universe for debugging
	context.count = len(my_stocks)
	All that glitters is not gold: Comparing backtest and out-of-sample performance of 800.000 trading algorithms

	“Past performance is no guarantee of future returns”. This cautionary message will certainly match the experience of many
	investors. When automated trading strategies are developed and evaluated using backtests on historical pricing data, there
	is always a tendency, intentional or not, to overfit to the past. As a result, strategies that show fantastic performance on
	historical data often flounder when deployed with real capital.

	Quantopian is an online platform that allows users to develop, backtest, and trade algorithmic investing strategies. By
	pooling all strategies developed on our platform we constructed a huge and unique data set of over 800.000 trading algorith
	ms. Although we do not have access to source code, we have returns and portfolio allocations as well as the time the
	<html>
	<table id="comparetable" class="clean">
	<tr>
	<td class="blank"> </td>
	<th>Machine Learning</th>
	<th>Frequentist<br>Statistics</th>
	<th>Probabilistic<br>Programming</th>
	</tr>
	<tr>
	<td class="rowTitle">Prediction</td>
	# After seeing Ex Machina I tweeted:
	# "So disappointed in #ExMachina -- the #Python wasn't PEP8 compatible..."
	# https://twitter.com/twiecki/status/599615426922938368
	# For the original code in the movie as well as what it does (it's kinda neat), see:
	# http://moviecode.tumblr.com/post/119171520870/in-the-movie-ex-machina-which-is-really-great

	# To put my code where my mouth is, I decided to make the code PEP8 compatible which you can see below.
	# I also did some minor cosmetic changes like changing i = i + 1 to i += 1
	# Other ideas for improvements? Comment below.