mmc2015的专栏

拼接原始数据：train_data = pd.read_csv('train.csv')test_data = pd.read_csv('test.csv')all_data = np.vstack((train_data.ix[:,1:-1], test_data.ix[:,1:-1]))numpy下的合并数组vstack和hstack函数：>>> a = np.ones((

美团团队的不少文章都不错http://tech.meituan.com/deep-understanding-of-ffm-principles-and-practices.html自己实现了一把，用python，速度慢于台大的libffm，但效果更好。几个注意点：1）label与loss function之间的关系。。。2）梯度计算时的 中间数据 预计

随机森林体现在随机上，台湾林老师讲了三种随机方式：1）样本bootstrap2）特征sample3）特征交叉组合看了sklearn的代码，实现了前两者，但没有第三种看了karpathy的代码（https://github.com/karpathy/Random-Forest-Matlab），没有bootstrap样本，feature组合部分写的也真是太“”随机“

http://www.analyticsvidhya.com/blog/2016/03/complete-guide-parameter-tuning-xgboost-with-codes-python/IntroductionIf things don’t go your way in predictive modeling, use XGboost.  

下面给出一个拟合多项式的例子：[python] view plaincopyimport pandas as pd    import numpy as np  import scipy as sp  import matplotlib.pyplot as plt      from sklearn.pipeline

http://dataunion.org/14892.htmlKaggle比赛源代码和讨论的收集整理。Algorithmic Trading Challenge40Solution whitepaper41.Solution thread30.Allstate Purchase Prediction Challenge7Rank 2 so

接上篇：def xgboost_pred(train,labels,test): params = {} params["objective"] = "reg:linear" params["eta"] = 0.005 params["min_child_weight"] = 6 params["subsample"] = 0.7 params["colsample

导入模块：import pandas as pdimport numpy as np #pandas依赖于numpyfrom sklearn import preprocessingimport xgboost as xgb常用功能简介：#load train and test train = pd.read_csv('train.csv', index_co

三张表；train_set.csv；test_set.csv；feature.csv。三张表通过object_id关联。import pandas as pdimport numpy as np# load training and test datasetstrain = pd.read_csv('../input/train_set.csv')test = pd.re

零、所有项目通用的：http://blog.youkuaiyun.com/mmc2015/article/details/46851245（数据集格式和预测器）http://blog.youkuaiyun.com/mmc2015/article/details/46852755（加载自己的原始数据）（适合文本分类问题的 整个语料库加载）http://blog.youkuaiyun.com/mmc2

print "Performing greedy feature selection..."score_hist = []N = 10good_features = set([])# Greedy feature selection loopwhile len(score_hist) score_hist[-2][0]: scores = [] for f in ran

# Hyperparameter selection loopscore_hist = []Cvals = [0.001, 0.003, 0.006, 0.01, 0.02, 0.03, 0.04, 0.05, 0.06, 0.1]for C in Cvals: model.C = C score = cv_loop(Xt, y, model, N) score_hi

http://machinelearningmastery.com/time-series-forecasting-long-short-term-memory-network-python/by Jason Brownlee on April 7, 2017 in Deep Learning0000The L