Advanced Data Structures: Designing an Optimal Count-Min Sketch

[@KK-dn6nc]

Hi, nice channel I like how you summarize it to its essentials

[@gabrielledatascience]

great explanation for bridging theory with practical coding application. Had a hard time visualizing what I learned in my lectures with what I needed for my assignments

[@ArunKumar-jk5pq]

is value of n is per some ‘time’? eg 1 million per day

[@niemasd]

n is per the entire lifetime of the Count-Min Sketch If you are trying to count e.g. the number of Netflix views each day, then you would create a new Count-Min Sketch every day and populate it throughout the day, and n would be the number of total views that entire day. If you are trying to count e.g. the number of Netflix views each week, then you would create a new Count-Min Sketch every week and populate it throughout the week, and n would be the total views that entire week So it depends on your specific use-case, but essentially n = total number of observations throughout the entire existence of the Count-Min Sketch, and the existence of the Count-Min Sketch can be some unit of time of interest (e.g. day, week, month, year, etc. depending on use-case)

[@ArunKumar-jk5pq]

@@niemasd Makes sense. Thanks!