bryangoodrich

#########################################
# Lesson 1 - R Basics
#
# Learning Objectives
# 1. R Data Types
# 2. Indexing
# 3. Boolean Logic and Filtering
# 4. Importing/Exporting
# 5. The R Environment
#########################################

bryangoodrich

################################################################################
# Author: Bryan Goodrich
# Created: Sometime 2012? Earlier?
#
# These functions allowed direct querying and mapping capabilities to the Bureau
# of Labor Statistics (BLS) Special Requests data through their FTP site. The 
# data were stored in fixed-width text files. 
#
# This code is non-functional in that the BLS no longer supports this FTP access.
# Additionally, the specific Local Area (LA) statistics utilized are also not

bryangoodrich

/*********************************************************************
 *  Compilation:  g++ grid_file.cpp -o grid_file.exe
 *  Execution:    grid_file.exe input.txt
 *  
 *  Input and rotate a grid
 *  
 *  This is a toy code snippet for handling informational
 *  input (integer size of grid), importing multiple
 *  columns, and exporting with formatted output. This
 *  example uses C++ approaches to file and string input

bryangoodrich

/*********************************************************************
 *  Compilation:  g++ grid.cpp -o grid.exe
 *  Execution:    grid.exe < input.txt
 *  
 *  Input and rotate a grid
 *  
 *  This is a toy code snippet for handling informational
 *  input (integer size of grid), importing multiple
 *  columns, and exporting with formatted output. Nothing
 *  amazing, but can provide a basis for this sort of 

bryangoodrich

/*********************************************************************
 *  Compilation:  g++ -std=c++11 map.cpp -o map.exe
 *  Execution:    map.exe < input.txt
 *  
 *  Output the contents of a sorted map
 *  
 *  This is a toy code snippet for using C++11 types to handle
 *  an input stream, store the incoming data columns by line, 
 *  and sort the output list according to length of key,
 *  with ties broken by string sort order (map default).

bryangoodrich

/*********************************************************************
 *  Compilation:  g++ -std=c++11 reverse.cpp -o reverse.exe
 *  Execution:    reverse.exe < input.txt
 *  
 *  Reverse an integer text input stream
 *  
 *  This is a toy code snippet for using C++11 types to handle
 *  an input stream, store the incoming data by line, and reverse
 *  it while printing to standard output.
 *  

bryangoodrich

# Twitter Topic Modeling Using R
# Author: Bryan Goodrich
# Date Created: February 13, 2015
# Last Modified: April 3, 2015
# 
# Use twitteR API to query Twitter, parse the search result, and
# perform a series of topic models for identifying potentially
# useful topics from your query content. This has applications for
# social media, research, or general curiosity
#

bryangoodrich

import array
class WeightedQuickUnion:
    def __init__(self, N):
        self.count = N
        self.id = array.array('i', range(N))
        self.size = array.array('i', [1] * N)
    
    def connected(self, p, q):
        return self.find(p) == self.find(q)

bryangoodrich

loadInput <- function() {
    structure(c(0.051267, -0.092742, -0.21371, -0.375, -0.51325, 
    -0.52477, -0.39804, -0.30588, 0.016705, 0.13191, 0.38537, 0.52938, 
    0.63882, 0.73675, 0.54666, 0.322, 0.16647, -0.046659, -0.17339, 
    -0.47869, -0.60541, -0.62846, -0.59389, -0.42108, -0.11578, 0.20104, 
    0.46601, 0.67339, -0.13882, -0.29435, -0.26555, -0.16187, -0.17339, 
    -0.28283, -0.36348, -0.30012, -0.23675, -0.06394, 0.062788, 0.22984, 
    0.2932, 0.48329, 0.64459, 0.46025, 0.6273, 0.57546, 0.72523, 
    0.22408, 0.44297, 0.322, 0.13767, -0.0063364, -0.092742, -0.20795, 
    -0.20795, -0.43836, -0.21947, -0.13882, 0.18376, 0.22408, 0.29896, 

bryangoodrich

gradientDescent <- function(X, y, initial_theta, method = "BFGS", ...) {
  m <- nrow(y)
  
  sigmoid      <- function(x) 1 / (1 + exp(-x))
  gradFunction <- function(theta) (1/m) * (t(X) %*% (sigmoid(X %*% theta)-y))
  costFunction <- function(theta) 
    (1/m) * (t(-y) %*% log(sigmoid(X %*% theta)) - t(1-y) %*% log(1 - sigmoid(X %*% theta)))  
  
  optim(initial_theta, costFunction, gradFunction, method = method, ...)  
}