data-mining-learn

Knowledge base for data mining

Affinity Analysis

Compare the different data mining algorithms for affinity analysis, especially apriori and eclat. Repo here.

CRISP-DM

Cross Industry Standard Practice for Data Mining. Try to implement the full steps.

Frequent Itemsets

Use frequent itemsets for food recommendation, create packages (food, drink and misc.), log the most frequently purchased and perform recommendations.

Mining high-utility itemsets in transaction database

Take a look into the algorithms and see how to apply it.

https://www.philippe-fournier-viger.com/spmf/mHUIMiner.php

Frecency

https://developer.mozilla.org/en-US/docs/Mozilla/Tech/Places/Frecency_algorithm

one way is to bucket the counter for each time the request hit, e.g. hourly

hour 24: 10 hour 23: 50 hour 22: 3 …

then, take the last N buckets to determine the ranking of the item. Another way is to add them in buckets with different range of data:

4 days bucket, weight: 100 14 days bucket: weight: 70 31 days bucket: weight 50 90 days bucket: weight 30 rest, 10:

The more frequent and recent ones will have a higher score. How can we do this with redis set?

Another easier way is just to check the usage in the last seven days. day 7 * 100 day 6 * 80 day 5 * 60 day 4 * 40 day 3 * 20 day 2 * 10 day 1 * 5 the sum of the scores is the recency score. Create fake data to check the distribution.

References

• https://www.janbasktraining.com/blog/what-is-data-mining/
• https://www.mssqltips.com/sqlservertutorial/4002/data-mining-algorithms-overview/
• https://searchdatamanagement.techtarget.com/definition/RFM-analysis
• https://stackoverflow.com/questions/43233300/sql-determine-recency-of-a-value
• https://www.deep-data-mining.com/2012/09/recency-frequency-monetary-rfm-analysis_9313.html
• http://www.silota.com/docs/recipes/sql-z-score.html