APRIORI AND FP GROWTH

mahajandiv

Hello.. Can anybody suggest me something for improving the performance of big data using association rules... for classification we have various methods to enhance as it is supervised learning ... but for unsupervised learning , i m not able to do it via association rules... yet know i have applied apriori and fp on my dataset get rules regarding that and then using log operator, i got the memory and time of my processes... 

Telcontar120

As I had mentioned earlier, one suggestion you should consider is sampling your dataset to a smaller size, at least to get things started.  This allows you to develop your process and figure out how to set things up without running into excessively long processing times or out of memory errors.  You can even begin to optimize your parameters with a smaller sample as well.  Eventually you may want to scale your solution up to "big data" levels (whatever that means for you) but it doesn't mean your entire project has to be conducted with all the data you have available.  There is a lot to be learned from fast prototyping with small datasets.