ltiao

• 安裝 easy_install, pip

  yum install python-setuptools
  easy_install pip

• 安裝 pexpect

ltiao

#!/usr/bin/python

import random

prob = {
  'a': (0, 0.4),
	'b': (0.4, 0.3),
	'c': (0.7, 0.2),
	'@': (0.9, 0.1)
}

ltiao

#!/usr/bin/python

import numpy as np
import matplotlib
matplotlib.use('TkAgg') # Required for Mac OSX
import matplotlib.pyplot as plt
import matplotlib.animation as animation

import logging.config
from settings import LOGGING

ltiao

1. CountVectorizer seems a bit dense and tries to do too much. This is not a bad thing but from a design viewpoint, it could benefit from making use of the Pipeline framework and separately define the preprocessor and tokenizer or even just analyzer as transformers in and of themselves and define CountVectorizer as a pipeline of these operatons. Should conform or make use of existing frameworks where possible rather than extending functionality to allow for exceptions.
2. Functions from the metric module such as confusion_matrix and classification_report ought to support cross-validation by perhaps allowing a cv parameter. This might be awkward given that we already have cross_val_score and learning_curve and those are supposed to be the functions which take care of cross validation scoring. It might make sense to use confusion_matrix as the scoring function for cross_val_score but the latter only accepts scoring functions which return a single value so that won't work. The other t

ltiao

ltiao

Core Sentiment Analysis Package

• Documentation, specifically docstrings. Set-up sphinx to organize documentation.
• Data loader need to refactored to respond gracefully to faults in the provided datasets.
• Helper function for concatenating train, dev and test data and providing the CV params (important/useful for learning_curve or other methods that only accept a single set of data with cross-validation parameters.)
• Turn the miscellaneous scripts into either package scripts or helper methods
• Clean up the IPython Notebooks and they can all be executed with "run all" and remove old/irrelevant ones
• Provide Cookbook with useful snippets
• Fix the package installer (setup.py) to support one-click install, so when the package is distributed, it will install everything required, including e.g. NLTK Copora, etc.
• Refactor code so all components are consistent. So far, data.py and feature_extraction.py are fully up-to-date and compliant with the design philosophy, but n

ltiao

ltiao

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% Structured General Purpose Assignment
% LaTeX Template
%
% This template has been downloaded from:
% http://www.latextemplates.com
%
% Original author:
% Ted Pavlic (http://www.tedpavlic.com)
%

ltiao

int sys_sbrk(__intptr_t change, void **retval) {
    struct addrspace *as = proc_getas();
	if (as == NULL) {
		return EFAULT;
	}

    vaddr_t heapbreak = as->as_heapvbase + as->as_heapsize;

    // we're not checking page alignment of 
    // change (amount) here or anything

ltiao

• lect07 (Slide 30-36) Review Classic File Organization methods
• lect16 (140513) 1:24: Top/bottom interrupt handling
• lect16 (140514) 0:10: Drivers are re-entrant
• lect16 (140514) Read up on the various I/O buffering in Tanenbaum
• lect17 (140520) Why favour I/O bound processes over CPU-bound processes?
• lect17 (140520) 00:31 Traditional UNIX scheduler (how priorities self-adjust with clock ticks etc.)
• lect17 (140520) 00:48 (Slide 62) Formula for scheduling
• lect17 (140520) 1:48 Why exactly do we favour Rate Monotonic scheduling other than that it is easier to implement (why is it easier to implement?)
• lect18 (Slide 12) What is the per-CPU private memory used for exactly?