{"code":1,"message":"","data":{"id":182247,"article":{"id":44631,"publicAccount":{"id":371,"name":"人工智能爱好者俱乐部","weixin":"aifans-club","intro":"","body":"","image":"http://ss.csdn.net/p?http://wx.qlogo.cn/mmhead/Q3auHgzwzM7UdrEE2iaZO6HBrUrb4sHDdmEIXSZONN243Ric6BoULWcw/0","href":null,"biz":"MzI2MzIwMDg4Ng==","category":{"id":2,"keyName":"AI","displayName":"人工智能","createdAt":1484122981000,"updatedAt":1497421384000,"status":1},"createdAt":1489375954000,"updatedAt":1489375954000,"status":1,"recommend":0,"qrcode":"http://mp.weixin.qq.com/mp/qrcode?scene=10000001&size=120&__biz=MzI2MzIwMDg4Ng==&mid=2651045534&idx=1&sn=48c4c97e82b1f4a9ce8a1176ba4880c1","articleCount":0,"viewCount":0},"category":{"id":2,"keyName":"AI","displayName":"人工智能","createdAt":1484122981000,"updatedAt":1497421384000,"status":1},"biz":"MzI2MzIwMDg4Ng==","mid":"2651046063","idx":2,"sn":"046c2e90d7cfb8de73f6810c454eff97","author":"","content":"","contentUrl":"http://mp.weixin.qq.com/s?__biz=MzI2MzIwMDg4Ng==&mid=2651046063&idx=2&sn=046c2e90d7cfb8de73f6810c454eff97&chksm=f148e9b6c63f60a031f68c100409e0d08916c177efb045e53996423aea59b9ed37537484ae65&scene=27#wechat_redirect","cover":"http:\\/\\/mmbiz.qpic.cn\\/mmbiz_jpg\\/pHfwfmSBicFVveUEgeB9NEOkh7icouu96EBejEibexMxSXZ8Br22U0epicm3GkaE8Hd4JH3ZJC9f1ibO7zBFibNsCKtg\\/0?wx_fmt=jpeg","digest":"作者：Gregory&nbsp;Piatetsky,&nbsp;KDnuggets本文中我们将对现有大数据的顶级工具做项测验：P","isMulti":0,"sourceUrl":"","title":"干货丨Python和R语言，谁更适用于Spark\\/Hadoop和深度学习","description":null,"datetime":1504566000000,"status":1,"viewCount":199,"praiseCount":17,"commentCount":0,"createdAt":1505096458000,"updatedAt":1505301799000,"origin":"proxy","recommend":0},"url":"http://mp.weixin.qq.com/s?__biz=MzI2MzIwMDg4Ng==&mid=2651046063&idx=2&sn=046c2e90d7cfb8de73f6810c454eff97&chksm=f148e9b6c63f60a031f68c100409e0d08916c177efb045e53996423aea59b9ed37537484ae65&scene=27#wechat_redirect","title":"干货丨Python和R语言，谁更适用于Spark\\/Hadoop和深度学习","source":"<p><img data-type=\"gif\" data-src=\"https://mmbiz.qpic.cn/mmbiz_gif/pHfwfmSBicFWR8nJxrL3dn3ibUsMSl5xSZu0rvwjDxX0cNTp8Ehqz3zQk7BI8rmSNHibrrNV0k7vCpYDVPZQ3EbMw/0?wx_fmt=gif\" class=\"\" data-ratio=\"0.15625\" data-w=\"640\" src=\"http://ss.csdn.net/p?https://mmbiz.qpic.cn/mmbiz_gif/pHfwfmSBicFWR8nJxrL3dn3ibUsMSl5xSZu0rvwjDxX0cNTp8Ehqz3zQk7BI8rmSNHibrrNV0k7vCpYDVPZQ3EbMw/640?wx_fmt=gif\"></p> \n<p><span style=\"color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; font-size: 16px; line-height: 28px;\">作者：Gregory Piatetsky, KDnuggets</span></p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">本文中我们将对现有<span style=\"color: rgb(0, 0, 255);\">大数据</span>的顶级工具做项测验：Python和R语言，谁更适用于<span style=\"color: rgb(0, 0, 255);\">Spark/Hadoop</span>和深度学习，并确定一个新兴的大数据深度学习生态系统。</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">上个月，我们对第18届KDnuggets软件的调查结果进行了一次报道：<span style=\"color: rgb(0, 0, 255);\">数据分析、数据科学、机器学习</span>中的新领导者、趋势和惊喜。</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">本文将更详细的考察哪些工具能够相互之间友好合作，哪些工具则兼容性较差。我们还发现了一个新兴的Python友好型工具生态系统，这些工具通常被应用在数据科学的两大前沿：大数据(Spark / Hadoop)和深度学习。</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">本文的末尾有一个匿名数据集的链接——欢迎读者对数据进行分析，然后将结果发布或者发送给我。</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">首先，让我们来看一下顶级工具之间的联系。</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">我们选取了投票超过500票的工具(今年有11项入选)。</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">衡量两个特征之间的关联程度有很多种方法，比如卡方分析或T检验，但此次我们仍然沿用了在2015年和2016年分析时采用的相对简单的方法。此处先定义一个“Lift”</p> \n<p style=\"text-align:center;padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\"><strong>Lift (X &amp; Y) = pct (X &amp; Y) / ( pct (X) * pct (Y) )</strong></p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">其中pct(X)表示选择X的用户百分比。</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">Lift(X&amp;Y)&gt; 1表示X&amp;Y 一起出现的频率比预设中两者相互独立要大;Lift=1表示X&amp;Y 一起出现的频率恰好等于预设中两者相互独立，Lift&lt;1表明X&amp;Y 一起出现的频率小于两者相互独立的情况(负相关)。</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">为了更直观的看到与1之间的差距，我们定义</p> \n<p style=\"text-align:center;padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\"><strong>Lift1 (X &amp; Y) = Lift (X &amp; Y) – 1</strong></p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">表1将排名前11的工具的Lift1值进行了两两比较，并筛选出关联度绝对值abs(Lift1) 大于15%的情况。</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">&nbsp;</p> \n<center style=\"color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; font-size: 16px; line-height: 28px; white-space: normal;\"> \n <img height=\"520\" data-src=\"http://mmbiz.qpic.cn/mmbiz_jpg/pHfwfmSBicFVveUEgeB9NEOkh7icouu96EhlXiaAGGo3icFUZXFyOW1hibblNGJXicKzDW9sMR2icm4LEEE8oy8ICAYgw/0?wx_fmt=jpeg\" width=\"806\" style=\"border: none; vertical-align: middle; display: inline; width: 660px; height: 425.806px;\" class=\"\" data-ratio=\"0.6451612903225806\" data-w=\"806\" data-type=\"jpeg\" src=\"http://ss.csdn.net/p?http://mmbiz.qpic.cn/mmbiz_jpg/pHfwfmSBicFVveUEgeB9NEOkh7icouu96EhlXiaAGGo3icFUZXFyOW1hibblNGJXicKzDW9sMR2icm4LEEE8oy8ICAYgw/640?wx_fmt=jpeg\"> \n</center> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">&nbsp;</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\"><strong>图1：数据科学、机器学习高级工具关联度表，2017</strong></p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">注：绿色表示正相关，红色表示负相关。</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">标签Lift1上文已解释;条形的宽度与Lift1的大小成正比。</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">我们注意到，Python不仅与Anaconda、Tensorflow和scikit- learn(不出所料)有显著的正相关，另外与Spark还具有显著正相关关系。</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">在比较流行的工具中，R语言相较于Python则关联性较弱。</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">除了Tableau之外，RapidMiner与其他顶级工具基本上都处于负相关关系，Excel用户也喜欢Tableau。而与Spark关系最密切的是Tensorflow 和scikit-learn。</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">Python聚类、Spark、Anaconda、Tensorflow和scikit- learn经常被一起使用，它们似乎形成了基于Python的大数据和深度学习生态系统的核心。</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\"><strong>Python vs R语言</strong></p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">接下来我们将研究使用Python或者R语言的前30个工具的亲和度。</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">用 with_Py(X)= %表示使用Python的工具X，with_R(X) % 表示使用R语言的工具X。可视化亲和度的过程中，我们采取一个非常简单的方式：Bias_Py_R(X) = log2(with_Py(X)/with_R(X)) ，若值为正则表明该工具更多使用了Python，当值为负则表明该工具更倾向于使用R语言。我们可以校正Python和R的相对频率，但由于它们在2017年的使用频率几乎相等，所以这种校正也是微不足道的。</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">&nbsp;</p> \n<center style=\"color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; font-size: 16px; line-height: 28px; white-space: normal;\"> \n <img height=\"708\" data-src=\"http://mmbiz.qpic.cn/mmbiz_jpg/pHfwfmSBicFVveUEgeB9NEOkh7icouu96EQOftOwWCFY9HoHSHYfj4sdqxGQA9JcaNNy7wgKAYGjEibnBMs6y9T2Q/0?wx_fmt=jpeg\" width=\"828\" style=\"border: none; vertical-align: middle; width: 660px; height: 564.348px;\" class=\"\" data-ratio=\"0.855072463768116\" data-w=\"828\" data-type=\"jpeg\" src=\"http://ss.csdn.net/p?http://mmbiz.qpic.cn/mmbiz_jpg/pHfwfmSBicFVveUEgeB9NEOkh7icouu96EQOftOwWCFY9HoHSHYfj4sdqxGQA9JcaNNy7wgKAYGjEibnBMs6y9T2Q/640?wx_fmt=jpeg\"> \n</center> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">&nbsp;</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\"><strong>表2：数据科学、机器学习前30位高级工具与Python vs R语言的关联度(2017)</strong></p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">条形图的长度为上文所定义的Bias_Py_R，条形图的高度与工具的受欢迎程度成正比。</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">我们注意到，与Python契合度较高的工具不仅包括我们预期的Scikit，PyCharm和Anaconda，而且还包括深度学习工具Keras和Tensorflow，特别是Spark和Scala。</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">与R语言契合度较高的工具包括SAS Base，Microsoft工具(预计自Microsoft购买Revolution Analytics后)，Weka和Tableau。</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">接下来，我们检查不同工具在大数据和深度学习中的效果</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">在KDnuggets 2017 Software Poll中，33%的受访者使用了Spark / Hadoop工具，32%使用了深度学习工具。完整的工具列表可在以下图表中查看。</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">对于每个工具X，我们计算与Spark / Hadoop工具(垂直轴)共同使用的频率以及Deep Learning工具(横轴)共同使用的频率。</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">&nbsp;</p> \n<center style=\"color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; font-size: 16px; line-height: 28px; white-space: normal;\"> \n <img height=\"621\" data-src=\"http://mmbiz.qpic.cn/mmbiz_jpg/pHfwfmSBicFVveUEgeB9NEOkh7icouu96ETnynbJLPzyqwktXIHgwzIiaKg1fRS354FEXQwibSicibwYicPDpXXUoz42Q/0?wx_fmt=jpeg\" width=\"748\" style=\"border: none; vertical-align: middle; width: 660px; height: 547.941px;\" class=\"\" data-ratio=\"0.8302139037433155\" data-w=\"748\" data-type=\"jpeg\" src=\"http://ss.csdn.net/p?http://mmbiz.qpic.cn/mmbiz_jpg/pHfwfmSBicFVveUEgeB9NEOkh7icouu96ETnynbJLPzyqwktXIHgwzIiaKg1fRS354FEXQwibSicibwYicPDpXXUoz42Q/640?wx_fmt=jpeg\"> \n</center> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">&nbsp;</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\"><strong>图3：深度学习vs Spark / Hadoop与顶尖数据科学，机器学习工具的亲和力(2017年)</strong></p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">圆形尺寸对应于各工具的使用份额，颜色与Python(蓝色)与R(橙色)相对应。</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">我们在图表的右上角注意到一组与Python相关的蓝色圆圈，包括scikit-learn，PyCharm，Anaconda，Java和Unix工具，这些工具更常用于Spark / Hadoop和深度学习工具。</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">这表明了一个Python-友好型的大数据/深度学习生态系统。</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">我们注意到Scala是与 “大数据”相关度最高的工具。</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">为了使图3更精确，它只包括至少获得200票的工具，及深度学习和Spark / Hadoop工具。请参阅下表1获取所有至少获得100票以上工具的更多详细信息。</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">&nbsp;</p> \n<center style=\"color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; font-size: 16px; line-height: 28px; white-space: normal;\"> \n <img height=\"821\" data-src=\"http://mmbiz.qpic.cn/mmbiz_jpg/pHfwfmSBicFVveUEgeB9NEOkh7icouu96EnqRiaQwwO6gZpIRXkdWKic86KAxB8OSaekqrhBgXVYT3Nfp1OcLR8icyw/0?wx_fmt=jpeg\" width=\"590\" style=\"border: none; vertical-align: middle; display: inline; height: 660px; width: 474.3px;\" class=\"\" data-ratio=\"1.3915254237288135\" data-w=\"590\" data-type=\"jpeg\" src=\"http://ss.csdn.net/p?http://mmbiz.qpic.cn/mmbiz_jpg/pHfwfmSBicFVveUEgeB9NEOkh7icouu96EnqRiaQwwO6gZpIRXkdWKic86KAxB8OSaekqrhBgXVYT3Nfp1OcLR8icyw/640?wx_fmt=jpeg\"> \n</center> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">&nbsp;</p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\"><strong>表1：深度学习vs Spark / Hadoop与顶尖数据科学，机器学习工具的亲和力(2017年)</strong></p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">下面是一个以CSV格式输入匿名的轮询数据的<span style=\"color: rgb(0, 0, 255);\">链接</span></p> \n<p style=\"padding-top: 8px; padding-bottom: 8px; word-wrap: break-word; font-size: 16px; line-height: 28px; color: rgb(85, 85, 85); font-family: 'Microsoft YaHei', Helvetica, Tahoma, Arial, sans-serif; white-space: normal;\">N：记录号（随机，记录不按投票顺序）<br>地区：usca：美国/加拿大、euro：欧洲/亚洲、Itam：拉丁美洲、afme：非洲/中东、aunz：澳大利亚/新西兰<br>Python：如果投票（最后一列）包含Python，则为1，否则为0<br>R语言：如果投票包含“R语言”，则为1，否则为0。我们使用“R语言”而不是R来简化正则表达式匹配<br>SQL语言：如果投票包含“SQL语言”，则为1，否则为0。<br>RapidMiner：如果投票包括RapidMiner，则为1，否则为0。<br>Excel：如果投票包含Excel，则为1，否则为0。<br>Spark：如果投票包括Spark，则为1，否则为0。<br>蟒蛇：如果投票包括蟒蛇，则为1，否则为0。<br>Tensorflow：如果投票包括Tensorflow，则为1，否则为0。<br>scikit-learn：如果投票包括scikit-learn，则为1，否则为0。<br>Tableau：如果投票包括Tableau，则为1，否则为0。<br>KNIME：如果投票包括KNIME，则为1，否则为0。<br>深度：如果投票包含深度，则为1，否则为0。<br>Spark / Hadoop：如果投票包括Spark / Hadoop，则为1，否则为0。<br>ntools：工具的数量<br>投票：选票列表，以分号分隔“;”</p> \n<p><img data-s=\"300,640\" data-type=\"jpeg\" data-src=\"https://mmbiz.qpic.cn/mmbiz_jpg/pHfwfmSBicFWY8b6F1aYhFoaQ0ibHSuU2Q2T1KK0TaQ2BhsPT47Zy9pJ6Zn2ZXYlpicvhw0icSoD0R0tNlMejgfJHQ/0?wx_fmt=jpeg\" class=\"\" data-ratio=\"0.9034749034749034\" data-w=\"518\" src=\"http://ss.csdn.net/p?https://mmbiz.qpic.cn/mmbiz_jpg/pHfwfmSBicFWY8b6F1aYhFoaQ0ibHSuU2Q2T1KK0TaQ2BhsPT47Zy9pJ6Zn2ZXYlpicvhw0icSoD0R0tNlMejgfJHQ/640?wx_fmt=jpeg\"></p> \n<p><br></p>","level":3,"createdAt":1505096458000,"updatedAt":1505301799000}}