CortanaAnalytics 0.0.4

安装

要安装，请使用pip:

pip install cortanaanalytics

您还可以直接从github repo获得开发版本：http://github.com/crwilcox/cortanaanalytics

开始

Cortana Analytics有许多不同的软件包。请看你感兴趣的图书馆的每个部分。

此外，您还需要来自azure数据市场的obtain an access key并订阅您希望使用的服务。

文本分析

https://datamarket.azure.com/dataset/amla/text-analytics

fromcortanaanalytics.textanalyticsimportTextAnalyticskey='1abCdEFGh/ijKlmN/opq234r56st/UvWXYZabCD7EF8='ta=TextAnalytics(key)score=ta.get_sentiment("hello world")scores=ta.get_sentiment_batch([{"Text":"hello world","Id":0},{"Text":"hello world again","Id":2}])

建议

https://datamarket.azure.com/dataset/amla/recommendations

fromcortanaanalytics.recommendationsimportRecommendationsemail='email@outlook.com'key='1abCdEFGh/ijKlmN/opq234r56st/UvWXYZabCD7EF8='rs=Recommendations(email,key)# create modelmodel_id=rs.create_model('groceries'+datetime.now().strftime('%Y%m%d%H%M%S'))# import item catalogcatalog_path=os.path.join('app','management','commands','catalog.csv')rs.import_file(model_id,catalog_path,Uris.import_catalog)# import usage informationtransactions_path=os.path.join('app','management','commands','transactions.csv')rs.import_file(model_id,transactions_path,Uris.import_usage)# build modelbuild_id=rs.build_fbt_model(model_id)status=rs.wait_for_build(model_id,build_id)ifstatus!=BuildStatus.success:print('Unsuccessful in building the model, failing now.')return# update model active build (not needed unless you are rebuilding)rs.update_model(model_id,None,build_id)print('Built a model. Model ID:{} Build ID:{}'.format(model_id,build_id))

异常检测

https://datamarket.azure.com/dataset/aml_labs/anomalydetection

fromcortanaanalytics.anomalydetectionimportAnomalyDetectionkey='1abCdEFGh/ijKlmN/opq234r56st/UvWXYZabCD7EF8='ad=AnomalyDetection(key)data=[(datetime(2014,9,21,11,5,0),3),(datetime(2014,9,21,11,10,0),9.09),(datetime(2014,9,21,11,15,0),0)]result=ad.score(test_data)

或者也可以使用字符串

fromcortanaanalytics.anomalydetectionimportAnomalyDetectionkey='1abCdEFGh/ijKlmN/opq234r56st/UvWXYZabCD7EF8='ad=AnomalyDetection(key)data="9/21/2014 11:05:00 AM=3;9/21/2014 11:10:00 AM=9.09;9/21/2014 11:15:00 AM=0;"params="SpikeDetector.TukeyThresh=3; SpikeDetector.ZscoreThresh=3"result=ad.score_raw(data,params)