研究设计与统计

订阅

MedSci主页 / 所有栏目 / 研究设计与统计

改进回归模型效果的六个小技巧

2015-12-18 法晓 医学统计分析精粹

回归模型技巧
关注

改进回归模型效果的六个小技巧 Six quick tips to improve your regression modeling 1. 拟合大量模型。 构思一系列可能的模型，从最简单到不可承受之繁。逐一尝试并及时舍弃，存留高效且相对简洁的模型。当然，需要提高模型拟合效率以实现此目的。 2. 使你的模型运算更快、更可信。 这貌似是一个计算方法问题，

改进回归模型效果的六个小技巧 Six quick tips to improve your regression modeling 1. 拟合大量模型。 构思一系列可能的模型，从最简单到不可承受之繁。逐一尝试并及时舍弃，存留高效且相对简洁的模型。当然，需要提高模型拟合效率以实现此目的。 2. 使你的模型运算更快、更可信。 这貌似是一个计算方法问题，但实则更多关乎统计。如果你可以更快地拟合模型，那么你就可以尝试更多的模型假设并进一步理解数据和模型。做到这一点，需要在数据准备或者模型复杂性上进一步动脑筋。 3. 对相关变量进行绘图。 对模型绘图 图形化展示数据是有趣的(^_^)，同样也是十分有效的。仅仅回归系数的列表并不能像绘图一样给你带来对模型的高度理解。这一优势相当明显，然而，在既往的教科书中较少提到。既往对绘图的使用，主要在探索原始数据和回归诊断的方面，而忽视了简单的作图就可以很好的帮助我们理解一个模型。 减少冗余绘图 你真的打算去绘制那些QQ图等统计软件包自动绘制的大量图片吗？你将如何使用这些绘制的图片呢？把它们忘记吧，专注到重要的事情上来。简单的法则：凡是你绘制的图片，都是你打算

版权声明：
本网站所有内容来源注明为“梅斯医学”或“MedSci原创”的文字、图片和音视频资料，版权均属于梅斯医学所有。非经授权，任何媒体、网站或个人不得转载，授权转载时须注明来源为“梅斯医学”。其它来源的文章系转载文章，或“梅斯号”自媒体发布的文章，仅系出于传递更多信息之目的，本站仅负责审核内容合规，其内容不代表本站立场，本站不负责内容的准确性和版权。如果存在侵权、或不希望被转载的媒体或个人可与我们联系，我们将立即进行删除处理。
在此留言
评论区 (5)
#插入话题
  1. [GetPortalCommentsPageByObjectIdResponse(id=187238, encodeId=b28718e23873, content=很实用，学习了。, beContent=null, objectType=article, channel=null, level=null, likeNumber=51, replyNumber=0, topicName=null, topicId=null, topicList=[], attachment=null, authenticateStatus=null, createdAvatar=https://wx.qlogo.cn/mmopen/Pw2csslt7CukWiabqiblnQAU2wVkTeANBK9R3rSMa9LcU72r7swR61YTJ7oZWL1MzhnJhRqMIsPZvCgHPwqmpYhDYKYK7dcwibS/0, createdBy=0cab1623974, createdName=hanchunhee, createdTime=Fri Apr 14 05:57:35 CST 2017, time=2017-04-14, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1981479, encodeId=b21519814e99f, content=<a href='/topic/show?id=02c1405214f' target=_blank style='color:#2F92EE;'>#回归模型#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=26, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=40521, encryptionId=02c1405214f, topicName=回归模型)], attachment=null, authenticateStatus=null, createdAvatar=, createdBy=ad7f282, createdName=yang0210, createdTime=Tue Jan 12 05:34:00 CST 2016, time=2016-01-12, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=48450, encodeId=679b48450b5, content=实用, beContent=null, objectType=article, channel=null, level=null, likeNumber=97, replyNumber=0, topicName=null, topicId=null, topicList=[], attachment=null, authenticateStatus=null, createdAvatar=, createdBy=45101626056, createdName=忠诚向上, createdTime=Sat Dec 26 17:16:00 CST 2015, time=2015-12-26, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=46617, encodeId=25bb4661edd, content=嗯嗯嗯, beContent=null, objectType=article, channel=null, level=null, likeNumber=116, replyNumber=0, topicName=null, topicId=null, topicList=[], attachment=null, authenticateStatus=null, createdAvatar=https://img.medsci.cn/20200609/ef6a57dbe7f1489dad7446b7c5217858/d5e4fe65996c4446a029b0f5d056abb1.jpg, createdBy=6b011518214, createdName=午夜星河, createdTime=Fri Dec 18 14:46:00 CST 2015, time=2015-12-18, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=46615, encodeId=5c124661533, content=赞, beContent=null, objectType=article, channel=null, level=null, likeNumber=81, replyNumber=0, topicName=null, topicId=null, topicList=[], attachment=null, authenticateStatus=null, createdAvatar=https://img.medsci.cn/20201205/e96b3d935daf4e258fba69d60175f0ab/b6d3ee4a3a414b8eafaf7c5e905937a2.jpg, createdBy=ad4b83208, createdName=vpssa, createdTime=Fri Dec 18 14:39:00 CST 2015, time=2015-12-18, status=1, ipAttribution=)]
    2017-04-14 hanchunhee

    很实用，学习了。

    0

  2. [GetPortalCommentsPageByObjectIdResponse(id=187238, encodeId=b28718e23873, content=很实用，学习了。, beContent=null, objectType=article, channel=null, level=null, likeNumber=51, replyNumber=0, topicName=null, topicId=null, topicList=[], attachment=null, authenticateStatus=null, createdAvatar=https://wx.qlogo.cn/mmopen/Pw2csslt7CukWiabqiblnQAU2wVkTeANBK9R3rSMa9LcU72r7swR61YTJ7oZWL1MzhnJhRqMIsPZvCgHPwqmpYhDYKYK7dcwibS/0, createdBy=0cab1623974, createdName=hanchunhee, createdTime=Fri Apr 14 05:57:35 CST 2017, time=2017-04-14, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1981479, encodeId=b21519814e99f, content=<a href='/topic/show?id=02c1405214f' target=_blank style='color:#2F92EE;'>#回归模型#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=26, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=40521, encryptionId=02c1405214f, topicName=回归模型)], attachment=null, authenticateStatus=null, createdAvatar=, createdBy=ad7f282, createdName=yang0210, createdTime=Tue Jan 12 05:34:00 CST 2016, time=2016-01-12, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=48450, encodeId=679b48450b5, content=实用, beContent=null, objectType=article, channel=null, level=null, likeNumber=97, replyNumber=0, topicName=null, topicId=null, topicList=[], attachment=null, authenticateStatus=null, createdAvatar=, createdBy=45101626056, createdName=忠诚向上, createdTime=Sat Dec 26 17:16:00 CST 2015, time=2015-12-26, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=46617, encodeId=25bb4661edd, content=嗯嗯嗯, beContent=null, objectType=article, channel=null, level=null, likeNumber=116, replyNumber=0, topicName=null, topicId=null, topicList=[], attachment=null, authenticateStatus=null, createdAvatar=https://img.medsci.cn/20200609/ef6a57dbe7f1489dad7446b7c5217858/d5e4fe65996c4446a029b0f5d056abb1.jpg, createdBy=6b011518214, createdName=午夜星河, createdTime=Fri Dec 18 14:46:00 CST 2015, time=2015-12-18, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=46615, encodeId=5c124661533, content=赞, beContent=null, objectType=article, channel=null, level=null, likeNumber=81, replyNumber=0, topicName=null, topicId=null, topicList=[], attachment=null, authenticateStatus=null, createdAvatar=https://img.medsci.cn/20201205/e96b3d935daf4e258fba69d60175f0ab/b6d3ee4a3a414b8eafaf7c5e905937a2.jpg, createdBy=ad4b83208, createdName=vpssa, createdTime=Fri Dec 18 14:39:00 CST 2015, time=2015-12-18, status=1, ipAttribution=)]
    2016-01-12 yang0210

    #回归模型#

    0

  3. [GetPortalCommentsPageByObjectIdResponse(id=187238, encodeId=b28718e23873, content=很实用，学习了。, beContent=null, objectType=article, channel=null, level=null, likeNumber=51, replyNumber=0, topicName=null, topicId=null, topicList=[], attachment=null, authenticateStatus=null, createdAvatar=https://wx.qlogo.cn/mmopen/Pw2csslt7CukWiabqiblnQAU2wVkTeANBK9R3rSMa9LcU72r7swR61YTJ7oZWL1MzhnJhRqMIsPZvCgHPwqmpYhDYKYK7dcwibS/0, createdBy=0cab1623974, createdName=hanchunhee, createdTime=Fri Apr 14 05:57:35 CST 2017, time=2017-04-14, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1981479, encodeId=b21519814e99f, content=<a href='/topic/show?id=02c1405214f' target=_blank style='color:#2F92EE;'>#回归模型#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=26, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=40521, encryptionId=02c1405214f, topicName=回归模型)], attachment=null, authenticateStatus=null, createdAvatar=, createdBy=ad7f282, createdName=yang0210, createdTime=Tue Jan 12 05:34:00 CST 2016, time=2016-01-12, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=48450, encodeId=679b48450b5, content=实用, beContent=null, objectType=article, channel=null, level=null, likeNumber=97, replyNumber=0, topicName=null, topicId=null, topicList=[], attachment=null, authenticateStatus=null, createdAvatar=, createdBy=45101626056, createdName=忠诚向上, createdTime=Sat Dec 26 17:16:00 CST 2015, time=2015-12-26, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=46617, encodeId=25bb4661edd, content=嗯嗯嗯, beContent=null, objectType=article, channel=null, level=null, likeNumber=116, replyNumber=0, topicName=null, topicId=null, topicList=[], attachment=null, authenticateStatus=null, createdAvatar=https://img.medsci.cn/20200609/ef6a57dbe7f1489dad7446b7c5217858/d5e4fe65996c4446a029b0f5d056abb1.jpg, createdBy=6b011518214, createdName=午夜星河, createdTime=Fri Dec 18 14:46:00 CST 2015, time=2015-12-18, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=46615, encodeId=5c124661533, content=赞, beContent=null, objectType=article, channel=null, level=null, likeNumber=81, replyNumber=0, topicName=null, topicId=null, topicList=[], attachment=null, authenticateStatus=null, createdAvatar=https://img.medsci.cn/20201205/e96b3d935daf4e258fba69d60175f0ab/b6d3ee4a3a414b8eafaf7c5e905937a2.jpg, createdBy=ad4b83208, createdName=vpssa, createdTime=Fri Dec 18 14:39:00 CST 2015, time=2015-12-18, status=1, ipAttribution=)]
    2015-12-26 忠诚向上

    实用

    0

  4. [GetPortalCommentsPageByObjectIdResponse(id=187238, encodeId=b28718e23873, content=很实用，学习了。, beContent=null, objectType=article, channel=null, level=null, likeNumber=51, replyNumber=0, topicName=null, topicId=null, topicList=[], attachment=null, authenticateStatus=null, createdAvatar=https://wx.qlogo.cn/mmopen/Pw2csslt7CukWiabqiblnQAU2wVkTeANBK9R3rSMa9LcU72r7swR61YTJ7oZWL1MzhnJhRqMIsPZvCgHPwqmpYhDYKYK7dcwibS/0, createdBy=0cab1623974, createdName=hanchunhee, createdTime=Fri Apr 14 05:57:35 CST 2017, time=2017-04-14, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1981479, encodeId=b21519814e99f, content=<a href='/topic/show?id=02c1405214f' target=_blank style='color:#2F92EE;'>#回归模型#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=26, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=40521, encryptionId=02c1405214f, topicName=回归模型)], attachment=null, authenticateStatus=null, createdAvatar=, createdBy=ad7f282, createdName=yang0210, createdTime=Tue Jan 12 05:34:00 CST 2016, time=2016-01-12, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=48450, encodeId=679b48450b5, content=实用, beContent=null, objectType=article, channel=null, level=null, likeNumber=97, replyNumber=0, topicName=null, topicId=null, topicList=[], attachment=null, authenticateStatus=null, createdAvatar=, createdBy=45101626056, createdName=忠诚向上, createdTime=Sat Dec 26 17:16:00 CST 2015, time=2015-12-26, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=46617, encodeId=25bb4661edd, content=嗯嗯嗯, beContent=null, objectType=article, channel=null, level=null, likeNumber=116, replyNumber=0, topicName=null, topicId=null, topicList=[], attachment=null, authenticateStatus=null, createdAvatar=https://img.medsci.cn/20200609/ef6a57dbe7f1489dad7446b7c5217858/d5e4fe65996c4446a029b0f5d056abb1.jpg, createdBy=6b011518214, createdName=午夜星河, createdTime=Fri Dec 18 14:46:00 CST 2015, time=2015-12-18, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=46615, encodeId=5c124661533, content=赞, beContent=null, objectType=article, channel=null, level=null, likeNumber=81, replyNumber=0, topicName=null, topicId=null, topicList=[], attachment=null, authenticateStatus=null, createdAvatar=https://img.medsci.cn/20201205/e96b3d935daf4e258fba69d60175f0ab/b6d3ee4a3a414b8eafaf7c5e905937a2.jpg, createdBy=ad4b83208, createdName=vpssa, createdTime=Fri Dec 18 14:39:00 CST 2015, time=2015-12-18, status=1, ipAttribution=)]
    2015-12-18 午夜星河

    嗯嗯嗯

    0

  5. [GetPortalCommentsPageByObjectIdResponse(id=187238, encodeId=b28718e23873, content=很实用，学习了。, beContent=null, objectType=article, channel=null, level=null, likeNumber=51, replyNumber=0, topicName=null, topicId=null, topicList=[], attachment=null, authenticateStatus=null, createdAvatar=https://wx.qlogo.cn/mmopen/Pw2csslt7CukWiabqiblnQAU2wVkTeANBK9R3rSMa9LcU72r7swR61YTJ7oZWL1MzhnJhRqMIsPZvCgHPwqmpYhDYKYK7dcwibS/0, createdBy=0cab1623974, createdName=hanchunhee, createdTime=Fri Apr 14 05:57:35 CST 2017, time=2017-04-14, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=1981479, encodeId=b21519814e99f, content=<a href='/topic/show?id=02c1405214f' target=_blank style='color:#2F92EE;'>#回归模型#</a>, beContent=null, objectType=article, channel=null, level=null, likeNumber=26, replyNumber=0, topicName=null, topicId=null, topicList=[TopicDto(id=40521, encryptionId=02c1405214f, topicName=回归模型)], attachment=null, authenticateStatus=null, createdAvatar=, createdBy=ad7f282, createdName=yang0210, createdTime=Tue Jan 12 05:34:00 CST 2016, time=2016-01-12, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=48450, encodeId=679b48450b5, content=实用, beContent=null, objectType=article, channel=null, level=null, likeNumber=97, replyNumber=0, topicName=null, topicId=null, topicList=[], attachment=null, authenticateStatus=null, createdAvatar=, createdBy=45101626056, createdName=忠诚向上, createdTime=Sat Dec 26 17:16:00 CST 2015, time=2015-12-26, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=46617, encodeId=25bb4661edd, content=嗯嗯嗯, beContent=null, objectType=article, channel=null, level=null, likeNumber=116, replyNumber=0, topicName=null, topicId=null, topicList=[], attachment=null, authenticateStatus=null, createdAvatar=https://img.medsci.cn/20200609/ef6a57dbe7f1489dad7446b7c5217858/d5e4fe65996c4446a029b0f5d056abb1.jpg, createdBy=6b011518214, createdName=午夜星河, createdTime=Fri Dec 18 14:46:00 CST 2015, time=2015-12-18, status=1, ipAttribution=), GetPortalCommentsPageByObjectIdResponse(id=46615, encodeId=5c124661533, content=赞, beContent=null, objectType=article, channel=null, level=null, likeNumber=81, replyNumber=0, topicName=null, topicId=null, topicList=[], attachment=null, authenticateStatus=null, createdAvatar=https://img.medsci.cn/20201205/e96b3d935daf4e258fba69d60175f0ab/b6d3ee4a3a414b8eafaf7c5e905937a2.jpg, createdBy=ad4b83208, createdName=vpssa, createdTime=Fri Dec 18 14:39:00 CST 2015, time=2015-12-18, status=1, ipAttribution=)]
    2015-12-18 vpssa

    赞

    0

相关资讯

BJJ：老年人半髋置换用骨水泥还是非骨水泥？

髋关节骨折是一种常见于老年患者的严重创伤，据估计，其发生率会在未来20年中持续增长。对于关节囊内的移位性股骨颈骨折常需要选择行半髋置换术，很多全国性的临床指南推荐骨水泥型半髋关节假体。而为了避免骨水泥带来的死亡率和并发症，有着严重合并症的病例往往对选择非骨水泥型假体。 过去的文献在这方面报道的结果一直存在争议。有少数患者甚至在术后短期就会发生死亡，但相关的研究往往又受到样本量的限制。来自

怀孕也挑时间？过短过长都不好

目前认为，怀孕间隔过短或过长都会影响胎儿发育。最近有研究表明，短期的怀孕间隔会增加新生儿患多动症的风险，但是，是否长期的怀孕间隔也会增加孤独症谱系障碍（ASD）的患病风险呢？暂时还未有研究人员对这一问题进行深入的研究。 针对这种情况，美国哥伦比亚大学梅尔曼公共卫生学院的Keely Cheslack-Postava博士进行了一项研究，研究结果显示，短期和长期的怀孕间隔都会增加新生儿ASD的患病风险

健康生活方式可降低慢性肾病患者心血管事件发生

　　美国伊利诺伊大学的Ana C. Ricardo博士在2012年肾脏周上报告的一项初步研究结果显示，健康生活方式可降低慢性肾病(CKD)患者心血管事件和死亡风险，但对肾脏事件风险无显著影响。 　　Ricardo博士指出，健康生活方式对普通人群影响的研究较为常见，但未见作为CKD患者不良结局预测指标的评价研究。此外，虽然也一些针对单一风险因素如吸烟与慢性肾病进展，以及锻炼与死亡率相关性研究

术中保护性机械通气可降低术后肺部并发症的发生

美国麻省总医院Karimd医师等的报告指出，术中保护性机械通气可降低术后肺部并发症的发生危险。潮气量≤8.5 mL/kg 预测体重(predicted body weight，PBW)和低平台压可使通气相关的主要术后肺部并发症的发生危险最小化。 每年有数百万的患者接受机械通气，然而至今也没有清晰的最佳术中通气策略共识来指导临床操作。虽然保护性机械通气在重症监护治疗病房中得到广泛应用，但在外科手术