(Venkat) Raghav, Rajagopalan raghavrv

New structure of tree builders/splitter/criterion

So right now what I've come up with is this:

A new criterion (only for mse) which will compute the stats for all the nodes. The criterion design will be as much similar as possible to the old one.

Two important data structures -

(node_stats numpy double array)

import os, time, gc, psutil
from pympler import tracker

import numpy as np
from sklearn.ensemble import ExtraTreesClassifier

tracker.memory_tracker = tracker.SummaryTracker()
def get_mem():
    return "{:.0f}MB".format(p.memory_info().rss / 1e6)

import os, time, gc, psutil
from pympler import tracker

import numpy as np
from sklearn.ensemble import RandomForestRegressor

tracker.memory_tracker = tracker.SummaryTracker()
def get_mem():
    return "{:.0f}MB".format(p.memory_info().rss / 1e6)

Test script

import os, time, gc, psutil
from pympler import tracker

import numpy as np
from sklearn.linear_model import LogisticRegression

tracker.memory_tracker = tracker.SummaryTracker()

Keybase proof

I hereby claim:

I am raghavrv on github.
I am raghavrv (https://keybase.io/raghavrv) on keybase.
I have a public key whose fingerprint is B127 1CD1 4F49 36C3 2FD0 2992 E301 E3AA 24EB 94A7

To claim this, I am signing this object:

	from types import FunctionType
	import inspect

	def undecorate(func):
	funcs = []

	# python 3.5
	if hasattr(func, '__wrapped__'):
	while hasattr(func, '__wrapped__'):
	func = func.__wrapped__

	%%cython

	# cython: wraparound=False
	# cython: boundscheck=False
	# cython: cdivision=True
	# cython: nonecheck=False
	# cython: profile=False

	cimport numpy as np
	import numpy as np

	from cpython cimport Py_INCREF, PyObject
	from libc.stdlib cimport malloc, free, calloc, realloc

	import numpy as np
	cimport numpy as np
	np.import_array()

	from sklearn.externals import joblib

	cdef struct SplitRecord:
	# Data to track sample splitting process
	# This structure also store the best split found so far
	SIZE_t feature # Which feature to split on.
	SIZE_t start
	SIZE_t end
	SIZE_t pos # Split samples array at the given position,
	# i.e. count of samples below threshold for feature.
	# pos is >= end if the node is a leaf.

	// ==UserScript==
	// @name URL in title
	// @version 0.1
	// @description try to take over the world!
	// @author Raghav RV
	// @updateURL https://gist.githubusercontent.com/raghavrv/eb2f44b43ef7d3287327a5fd5e3ee205/raw/url_in_title.js
	// @downloadURL https://gist.githubusercontent.com/raghavrv/eb2f44b43ef7d3287327a5fd5e3ee205/raw/url_in_title.js
	// @include http:///
	// @include https:///
	// @grant unsafeWindow