0% found this document useful (0 votes)
85 views6 pages

(New) (New) ML KNN Introduction Handwritten Notes

rwe safd

Uploaded by

josoh70168
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
0% found this document useful (0 votes)
85 views6 pages

(New) (New) ML KNN Introduction Handwritten Notes

rwe safd

Uploaded by

josoh70168
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
You are on page 1/ 6
Le Mosty used fow ciasstfication, a . Le non- pavametsic algo: Not make any 2 oe” undestying data. Up lazy leasnek_algo- Non learn teoaining immediatly, instead stove dataset 2 at classi time, pecfoor action. Nee knn ab tvaining phase Just stoswe da on new data, classify data into category 3A preotctaa sss e @Q ko wew ata. x2 Testing @ Manhbatteo Wweateulate Aist. by sv bet". comcesponding featuse values. Le Suitable aata with atte have Aiffesmnt unt aiffesence be bee? Ay Calculate summing absolute coovdinates of Bowes tn otdims. Aalor,y) = ee 7 ae @® Minkowski Distance — CS Genewaliea” of woth Fuclideno Aietaa p=?) & Mankatten Distence (pr 4). Pits when pas => Reduces to pies Euclidean dist. CE acee Ak! with best validation @ dad *k' value — oo” 4 lw choose oda ‘k' values to avoid te Le sf th! even: vesult in ambiguous @© ure of tThumb— aoe Wie = Asound squam soot of num ef data lw As data tes = smallew ‘hk’ betters © asa Seasch — G rexfoom geid seach oves wange of *e e Xe Pesfoum cevoss- valiclatinn & find best Se oe aos sss © Lonsides Data Chasactesistics — OG smaies *k' captiwe jodal Pasteaaay Xe Noisy dataset => Lasges 'K" © Bias — Vawlance Azedssit ol Le Smattews tk! 5 Ly Lavges ‘Kk’ + ry Should balance ‘: a smooth, Example: fseom below dataset. Pred x= (mMath=6, cs=8) fou = Math | es | Resuit 3 | Fatt é ve Pass 1 Fass fe eles Fail iS 8 Pass Find Euclidean dist- of al! Points? 0: fan) Seer vile = Xa, 17 ! oe - xe | | Muidemont (e-9)) 2 Jeg = 5-32) [cess + ¢ee7)> = VN = oO ©. (6-77 +¢a-28)" ice) to * Fass’. a Poin*® (6,9) belongs @ data Preprocessing- can use NN to missing values cla imputation method. -_ 2 ee @ patteso Recogniton— TtTeain KHN model with Mutat dataset & theo pectowm evatuation Pp @ sesults too high accuracy. : © Recommendation Engines— Assign each pacticulas group & then pwovide tecomm based on group's prefewences, © Umitations of KN — © Does not scale — lazy seawnes ? comput" & mom stowage. Time-censuming & wscum, Figs @ Cowse of Dimenstonality- vasa to) sien data points when too bigh Aimenstonaltty, © rzone to Ovesfitting— Lesa to overs Freting we2 of cucse OF Aimensionality. ; So, use featucr selection & AYemensionaly peduction.

You might also like