If you have ever used Linkedin Learning or atleast seen some recommended courses on your news feed, you might have wondered how a particular course comes up in your suggestion.
There are three steps or process layers that determine which course will be shown to you
- 𝐑𝐞𝐬𝐩𝐨𝐧𝐬𝐞 𝐏𝐫𝐞𝐝𝐢𝐜𝐭𝐢𝐨𝐧 ✅️
Predicting your response uses 3 main factors or data points to train the AI model.
• Profile Features - Your skills & Industry
• Course Metadata - Course difficulty, skills & category
• Historical engagements - Clicks, bookmarks etc.
- 𝐂𝐨𝐥𝐥𝐚𝐛𝐨𝐫𝐚𝐭𝐢𝐯𝐞 𝐅𝐢𝐥𝐭𝐞𝐫𝐢𝐧𝐠 💠
Here, direct data is not used. Rather similarity is measured.
• which courses are chosen by users who have similar profile with you ; and
• which kind of users chose a particular course
By mapping this from both sides they get a collaborative filtered output.
And these models are continuously updated each second.
- 𝐁𝐥𝐞𝐧𝐝𝐢𝐧𝐠 🔃
Once the output from both models are available, they are blended to generate the most favorable result for each user.
🤔 𝐖𝐡𝐚𝐭 𝐛𝐞𝐧𝐞𝐟𝐢𝐭 𝐝𝐨 𝐲𝐨𝐮 𝐠𝐞𝐭 𝐛𝐲 𝐤𝐧𝐨𝐰𝐢𝐧𝐠 𝐭𝐡𝐢𝐬?
Recommendations could show up topics that you might have never thought of learning but are actually relevant to you.
Hence keep all your input points clean and relevant, so you get the best suggestions from the AI. 🙂
