BigQuery ML を使用した ML モデルの作成: チャレンジラボ のレビュー

The instructions in step 3 are ambiguous. There's no way to know whether the model was created according to the instructions, or whether the evaluation step doesn't pass the checker. Even after 7 separate attempts to run step 3, I can't pass the checker. ## Task 3. Improve model performance with Feature Engineering and Evaluate the model to see if there is better predictive power """ In this task, use dataset features that may help the customer_classification_model model better understand the relationship between a visitor's first session and the likelihood that they purchase on a subsequent visit. Now add some new features and create a second machine learning model called improved_customer_classification_model. How far the visitor got in the checkout process on their first visit Where the visitor came from(traffic source: organic search, referring site, etc..) Device category(mobile, tablet, desktop) Geographic information(country) Now, evaluate the newly created model improved_customer_classification_model to see if there is better predictive power than customer_classification_model. """ CREATE OR REPLACE MODEL `ecommerce.improved_customer_classification_model` OPTIONS ( model_type='logistic_reg', labels = ['will_buy_on_return_visit'] ) AS #standardSQL SELECT * EXCEPT(fullVisitorId) FROM # features (SELECT fullVisitorId, IFNULL(totals.bounces, 0) AS bounces, IFNULL(totals.timeOnSite, 0) AS time_on_site, MAX(CAST(h.eCommerceAction.action_type AS INT64)) AS latest_ecommerce_progress, IFNULL(geoNetwork.country, "") AS country, trafficSource.source as traffic_source, device.deviceCategory as device_category, IFNULL(geoNetwork.country, "") AS the_country FROM `data-to-insights.ecommerce.web_analytics`, UNNEST(hits) as h WHERE totals.newVisits = 1 AND date BETWEEN '20160801' AND '20170430' # train on first 9 months GROUP BY fullVisitorId, bounces, time_on_site, the_country, traffic_source, device_category) JOIN (SELECT fullvisitorid, IF(COUNTIF(totals.transactions > 0 AND totals.newVisits IS NULL) > 0, 1, 0) AS will_buy_on_return_visit FROM `data-to-insights.ecommerce.web_analytics` GROUP BY fullvisitorid) USING (fullVisitorId); # 3.5 SELECT roc_auc, CASE WHEN roc_auc > .9 THEN 'good' WHEN roc_auc > .8 THEN 'fair' WHEN roc_auc > .7 THEN 'decent' WHEN roc_auc > .6 THEN 'not great' ELSE 'poor' END AS model_quality FROM ML.EVALUATE(MODEL ecommerce.improved_customer_classification_model, ( SELECT * EXCEPT(fullVisitorId) FROM # features (SELECT fullVisitorId, IFNULL(totals.bounces, 0) AS bounces, IFNULL(totals.timeOnSite, 0) AS time_on_site, MAX(CAST(h.eCommerceAction.action_type AS INT64)) AS latest_ecommerce_progress, IFNULL(geoNetwork.country, "") AS country, trafficSource.source as traffic_source, device.deviceCategory as device_category, IFNULL(geoNetwork.country, "") AS the_country FROM `data-to-insights.ecommerce.web_analytics`, UNNEST(hits) as h WHERE totals.newVisits = 1 AND date BETWEEN '20160801' AND '20170430' # train on first 9 months GROUP BY fullVisitorId, bounces, time_on_site, the_country, traffic_source, device_category) JOIN (SELECT fullvisitorid, IF(COUNTIF(totals.transactions > 0 AND totals.newVisits IS NULL) > 0, 1, 0) AS will_buy_on_return_visit FROM `data-to-insights.ecommerce.web_analytics` GROUP BY fullvisitorid) USING (fullVisitorId) ));

Meyer Jennifer · 28日前にレビュー済み

Tang Karwoo · 28日前にレビュー済み

BUG in task 2 and 3. Cant be completed in some cases

L Johan · 28日前にレビュー済み

Augusto Dos Santos Victor · 28日前にレビュー済み

Tang Karwoo · 28日前にレビュー済み

Holy-Mafia Da · 28日前にレビュー済み

Tang Karwoo · 28日前にレビュー済み

Augusto Dos Santos Victor · 28日前にレビュー済み

Lopez Reyna · 28日前にレビュー済み

P V S K P Phani Kumar · 28日前にレビュー済み

. Anand Ravi · 28日前にレビュー済み

in this lab, the second test case is passing whatever we do, even after evaluating the model, it is not seen as pass. Could you pls check if there is any issue in that ?

Yerasi Subhash · 28日前にレビュー済み

Phan Vinh · 28日前にレビュー済み

Birsan Marius-Dumitru · 28日前にレビュー済み

Tomar Monica · 28日前にレビュー済み

Antony Ashmore Lee · 28日前にレビュー済み

Antony Ashmore Lee · 28日前にレビュー済み

Mysorematt Nandeesh S · 28日前にレビュー済み

. Nandini Gajjala · 28日前にレビュー済み

Gupta Devansh · 28日前にレビュー済み

Singh Harshit · 28日前にレビュー済み

lab is wired and can not pass

Zhang Hui · 28日前にレビュー済み

very bad, i dont know why i cannot pass task 2 and 3, while i have done dozens of methods to evaluate the models, if i'm really doing something incorect, the lab checks should tell me something useful , rather than keeping giving vague msg for learner to guess, it's a waste of time..

YAO WEI TAO VICTOR · 28日前にレビュー済み

Takkar Minkush · 28日前にレビュー済み

Mahadasa Chaitanya · 28日前にレビュー済み