• Supervised Learning:
• evaluating performance-Classificaion error matrix
• evaluating performance- Regression Error Matrix
• Machine Learning With Python
• yt-campusx- accuracy - recoil, precision

precion: jitni chizo ko apne positive predict kiya hai, unmese kitni positive hai

• true positive divide kardo ge jitno ko positive bola hai
• spam example

recal: jitne sache mai positive hai unmaise kitne logo ko cancer detect kiya gaya

• cancer example

• yaha tumhare modl ko kya hai, type1 jyada problematic hai, ya type 2 jyada problematic hai, to us hisab se vo kar lega

2. dono ke bich mai trade off hota hai, recal ^ → precision kam hoga
3. agar precision ^ → recal kam hoga
4. f1 score hamesha lower wali side rahega
5. agar dono chizo ko dekhna hai to f1 score use hota hai