spark | spark 机器学习chapter3 数据的获取、处理与准备
spark | spark 机器学习chapter3 数据的获取、处理与准备
阅读spark机器学习这本书来学习在spark上做机器学习
注意：数据集是电影评分等数据，下载链接：http://files.grouplens.org/datasets/movielens/ml-100k.zip
数据集包括：用户属性文件、电影元素、用户对电影的评级 １、将数据解压到某个目录下，并切换到该目录
unzip ml-100k.zip
cd ml-100k２、查看上述三种数据
用户
电影
评分
３、启动python，分析数据
启动
/home/hadoop/spark/bin/pyspark４、读数据
from pyspark import SparkContext
user_data = sc.textFile("u.user")
user_data.first()u’1|24|M|technician|85711’
５、基本的分析
＃分割数据，函数split
user_fields=user_data.map(lambda line:line.split("|"))
#用户数量
num_users=user_fields.map(lambda fields:fields[0]).count()
＃性别数
num_genders = user_fields.map(lambda fields:fields[2]).distinct().count()
＃职业种数
num_occupations = user_fields.map(lambda fields:fields[3]).distinct().count()
＃其他
num_zipcodes=user_fields.map(lambda fields:fields[4]).distinct().count()
＃结果打印
print "Users:%d ,genders:%d ,occupations:%d ,ZIP codes:%d" % (num_users,num_genders,num_occupations,num_zipcodes)Users:943 ,genders:2 ,occupations:21 ,ZIP codes:795
6、画图 对ages这个属性做直方图。 由于在终端下没法画图，这里给出代码
ages = user_fields.map(lambda x: int(x[1])).collect()
hist(ages, bins=20, color='lightblue', normed=True)
fig = matplotlib.pyplot.gcf()
fig.set_size_inches(16, 10)７、统计职业的种类和数量
import numpy as np
count_by_occupation = user_fields.map(lambda fields:(fields[3],1)).reduceByKey(lambda x,y:x+y).collect()
x_axis1 = np.array([c[0] for c in count_by_occupation])
y_axis1 = np.array([c[1] for c in count_by_occupation])打印出结果
print x_axis1[u’administrator’ u’retired’ u’lawyer’ u’none’ u’student’ u’technician’ u’programmer’ u’salesman’ u’homemaker’ u’executive’ u’doctor’ u’entertainment’ u’marketing’ u’writer’ u’scientist’ u’educator’ u’healthcare’ u’librarian’ u’artist’ u’other’ u’engineer’]
print y_axis1[ 79 14 12 9 196 27 66 12 7 32 7 18 26 45 31 95 16 51 28 105 67]
y_axis = y_axis1[np.argsort(y_axis1)]array([ 7, 7, 9, 12, 12, 14, 16, 18, 26, 27, 28, 31, 32, 45, 51, 66, 67, 79, 95, 105, 196])
np.argsort() : 得到升序的下标
画图：
pos = np.arange(len(x_axis))
width = 1.0
ax = plt.axes()
ax.set_xticks(pos + (width / 2))
ax.set_xticklabels(x_axis)
plt.bar(pos, y_axis, width, color='lightblue')
plt.xticks(rotation=30)
fig = matplotlib.pyplot.gcf()
fig.set_size_inches(16, 10)计算各个值出现不同次数的方法：
count_by_occupation2 = user_fields.map(lambda fields: fields[3]).countByValue()
print "Map-reduce approach:"
print dict(count_by_occupation2)
print ""
print "countByValue approach:"
print dict(count_by_occupation)Map-reduce approach print(dict(count_by_occupation2)) {u’administrator’: 79, u’retired’: 14, u’lawyer’: 12, u’healthcare’: 16, u’marketing’: 26, u’executive’: 32, u’scientist’: 31, u’student’: 196, u’technician’: 27, u’librarian’: 51, u’programmer’: 66, u’salesman’: 12, u’homemaker’: 7, u’engineer’: 67, u’none’: 9, u’doctor’: 7, u’writer’: 45, u’entertainment’: 18, u’other’: 105, u’educator’: 95, u’artist’: 28}
countByValue approach {u’administrator’: 79, u’writer’: 45, u’retired’: 14, u’lawyer’: 12, u’doctor’: 7, u’marketing’: 26, u’executive’: 32, u’none’: 9, u’entertainment’: 18, u’healthcare’: 16, u’scientist’: 31, u’student’: 196, u’educator’: 95, u’technician’: 27, u’librarian’: 51, u’programmer’: 66, u’artist’: 28, u’salesman’: 12, u’other’: 105, u’homemaker’: 7, u’engineer’: 67}
探索电影数据 解析电影分类数据的特征
读数据和查看数据
读数据
movie_data = sc.textFile("u.item")查看数据
＃第一行
print movie_data.first()1|Toy Story (1995)|01-Jan-1995||http://us.imdb.com/M/title-exact?Toy%20Story%20(1995)|0|0|0|1|1|1|0|0|0|0|0|0|0|0|0|0|0|0|0
电影总数
print "Movies:%d" % num_moviesMovies:1682
对电影发型的时间做处理 先过虑掉缺失值，定义函数，缺失值取为1900
def convert_year(x):
try:
return int(x[-4:])
except:
return 1900 第３列为时间，格式为：01-Jan-1995 ，－４：得到年
数据处理
movie_fields = movie_data.map(lambda lines: lines.split("|"))
years = movie_fields.map(lambda fields: fields[2]).map(lambda x: convert_year(x))过滤掉为1900的
years_filtered = years.filter(lambda x: x != 1900)计算电影的年龄，该数据发生在1998年，要得到发行时间，需要1998减去时间
movie_ages = years_filtered.map(lambda yr: 1998-yr).countByValue()
values = movie_ages.values()
bins = movie_ages.keys()
print values
print bins[65, 286, 355, 219, 214, 126, 37, 22, 24, 15, 11, 13, 15, 7, 8, 5, 13, 12, 8, 9, 4, 4, 5, 6, 8, 4, 3, 7, 3, 4, 6, 5, 2, 5, 2, 6, 5, 3, 5, 4, 9, 8, 4, 5, 7, 2, 3, 5, 7, 4, 3, 5, 5, 4, 5, 4, 2, 5, 8, 7, 3, 4, 2, 4, 4, 2, 1, 1, 1, 1, 1]
[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 72, 76]
画图
hist(values, bins=bins, color='lightblue', normed=True)
fig = matplotlib.pyplot.gcf()
fig.set_size_inches(16,10)探索评级数据
读数据、数据量
>>> #评级数据
...
rating_data = sc.textFile("u.data")
print rating_data.first()
196 242 3 881250949
num_ratings = rating_data.count()
print "Ratings: %d" % num_ratings
Ratings: 100000196 242 3 881250949 Ratings: 100000 总共10万条数据
一些基本统计
#数据分割
rating_data1 = rating_data.map(lambda line:line.split("\t"))
#评分
ratings = rating_data1.map(lambda fields:int(fields[2]))
#最高得分
max_rating = ratings.reduce(lambda x,y:max(x,y))
#最低得分
min_rating = ratings.reduce(lambda x,y:min(x,y))
#评价得分
mean_rating = ratings.reduce(lambda x,y:x+y) / num_ratings
#中位数
median_rating = np.median(ratings.collect())
#平均每个用户打分数
ratings_per_user = num_ratings / num_users
＃平均每部电影有多少评分
ratings_per_movie = num_ratings / num_movies
print "Min ratings: %d" % min_rating
print "Max rating: %d" % max_rating
print "Average rating: %2.2f" % mean_rating
print "Median rating: %d" % median_rating
print "Average # of rating per user:%2.2f" % ratings_per_user
print "Average # of ratings per movie: %2.2f" % ratings_per_movieMin ratings: 1 Max rating: 5 Average rating: 3.00 Median rating: 4 Average # of rating per user:106.00 Average # of ratings per movie: 59.00
类似功能的函数stats()
ratings.stats()(count: 100000, mean: 3.52986, stdev: 1.12566797076, max: 5.0, min: 1.0)
计算每个用户打分次数：
user_ratings_grouped = rating_data.map(lambda fields:(int(fields[0]),int(fields[2]))).groupByKey()
user_ratings_byuser = user_ratings_grouped.map(lambda (k,v):(k,int(len(v))))
user_ratings_byuser.take(5) ＃这里在spark2.1下报错，后续探究绘图
user_ratings_byuser_local = user_ratings_byuser.map(lambda (k, v): v).collect()
hist(user_ratings_byuser_local, bins=200, color='lightblue', normed=True)
fig = matplotlib.pyplot.gcf()
fig.set_size_inches(16,10)腾讯云开发者
