منتديات نورة الحربي

 

 


العودة   منتديات نورة الحربي > مواضيع منقولة من مواقع اخرى2

إضافة رد
 
أدوات الموضوع انواع عرض الموضوع
  #1  
قديم 11-15-2012, 03:31 PM
rss
Guest
 
المشاركات: n/a
افتراضي Behind the Scenes: Twitter, Part 3 - A win for simple

 

This is the final Part of a three Part ****** on how we use Twitter as a support channel. In the first part, I described the tools we use to manage Twitter; in the second part, we built a model to separate tweets into those that need an immediate response or **t.

In the arc of a three Part ******, the final Part is supposed to be either a story of triumph or an object lesson.

 


The triumphant story would be about how we implemented the model we built previously in production, integrated it with our Rails-based Twitter client, and saw massive quantifiable improvements resulting from it. I would look smart, competent, and impactful.

 


The object lesson is that sometimes practical concerns win out over a neat tech**logical solution, and that’s the story here.

Sometimes good isn’t good e**ugh

The model we built had a false positive rate of about 7%. That’s fair, and in many applications, that would be just fine. In our case, we want to be very confident we’re **t missing important tweets from people that need help. Practically, that means that someone would have to check the classification results occasionally to find the handful of tweets that do need an immediate response that slipped through.
After talking to the team, it became pretty clear that checking for mis-classified tweets would be more work than just handling the full, unclassified feed with the manual keyword filtering we have been using. This is a great example of a case where absolute numbers are more important than percentages: while the percentage impact in terms of filtering out less urgent tweets would be significant, the actual practical impact is much more muted because we’ve optimized the tool to handle tweets quickly.
Part of the reason why we’re able to get away with keyword filtering rather than something more sophisticated is because of just how accurate it is with essentially ** false positives. There’s actually a surprising amount of duplication in tweets—excluding retweets, the last 10,000 tweets we’ve indexed have only 7,200 unique bodies among them. That means that when a person looks at the first tweet using a phrase they can instantly identify that there’s a keyword that’s going to reoccur (for example, as soon as I started this ******, we added “Behind the Scenes: Twitter” to the keyword list) and add it to the keyword list.

Most of the benefit with little effort

Rather than putting the predictive model into production, I switched our keyword filters from a hardcoded array to a Redis set, and added a form to add new keywords to list in a manner of seconds. This took five minutes to add rather than the five hours it might have taken to hook up a live predictive model.
I haven’t forgotten about the model. It’s in a proverbial desk drawer, and in a few months I’ll revisit it. I might have some new ideas of things to try to improve the false positive rate, and we’ll have a bigger training set as well. Heroics and fancy tech**logy might **t have won today, but the important thing is to meet the business’s needs in as efficient a manner as possible.

Behind Scenes: Twitter, Part simple Behind Scenes: Twitter, Part simple
Behind Scenes: Twitter, Part simple

أكثر...


Behind the Scenes: Twitter< Part 3 - A win for simple

رد مع اقتباس
إضافة رد

مواقع النشر (المفضلة)

أدوات الموضوع
انواع عرض الموضوع

تعليمات المشاركة
لا تستطيع إضافة مواضيع جديدة
لا تستطيع الرد على المواضيع
لا تستطيع إرفاق ملفات
لا تستطيع تعديل مشاركاتك

BB code is متاحة
كود [IMG] متاحة
كود HTML معطلة

الانتقال السريع

Bookmark and Share

RSS RSS 2.0 XML MAP HTML

الساعة الآن 05:09 PM

أقسام المنتدى

الشريعة و الحياة | المنتدى العام | مـنـتـدى الـعـرب الـمـسافـرون | أزياء - فساتين - عبايات - ملابس نسائية - موضة | مكياج - ميك اب - عطورات - تسريحات شعر - العناية بالبشرة | װ.. أطآيبُ ״ شّهية ]●ะ | اعشاب طبية - الطب البديل - تغذية - رجيم - العناية بالجسم | ديكور - اثاث - غرف نوم - اكسسوارات منزلية | صور - غرائب - كاريكاتير | {.. YouTube..} | منتدى الاسئله والاستفسارات والطلبات | مواضيع منقولة من مواقع اخرى | اخبار التقنية | مواضيع منقولة من مواقع اخرى2 | منتدى تغذيات | منتدى تغذيات الكلي | قِصصْ الأنْبِيَاء | قِسمْ الرَّسُول الكَرِيمْ والصَّحَابة الكِرامْ | القِسمْ الإِسْلاَمِي العَامْ | قِسمْ المَواضِيْع العَامَّة | قِسمْ الشِعرْ والشُعَرَاء | قِسمْ الخَوَاطِر المَنقُولَة | قِسمْ الصُوَرْ | قِسمْ القِصصْ والرِوَايَات | قِسمْ حَوَّاءْ | قِسمْ الطِب والصِحَّة | قِسمْ الصَوتِيَّات والمَرئيَّات الإسْلاميَّة | قِسمْ مَطبَخ صَمْت الوَدَاعْ | قِسمْ الدِيكورْ | قِسمْ كٌرَة القََدَمْ العَرَبِيَّة والعَالمِيَّة | قِسمْ المكْياجْ والإكْسسْوارَات | قِسمْ المَاسِنجرْ | قِسمْ الِسِياحَة والسَفرْ | قِسمْ أَفْلاَمْ الكَرتُونْ | قِسمْ الفِيدْيو المُتَنَوعْ | منتديات الرياضة | اخبار | مواضيع منقولة من مواقع اخرى3 | موقع اجنبي | كوكو فرنسا | كوكو هندية | كوكو روسي | كوكو دنمارك | كوكو ياباني | اخر اخبار العولمة | عالم الحياة الزوجية | عرض أحدث عمليات البحث الشائعة |



Powered by vBulletin® Version 3.8.8 Beta 1
Copyright ©2000 - 2017, KingFromEgypt Ltd Elassal
Adsense Management by Losha
This Forum used Arshfny Mod by islam servant