I've been hunting for a more suitable languages do do pattern mining & clustering work in for a while. Most likely such a language would be declarative and take care of alot of what I currently write myself in procedural land (backtracking etc).
Currently using:
PHP to interface with an Elasticsearch cluster. Most of the work is in the Elasticsearch queries with PHP being used for educated guesses and backtracking.
Clustering can either be based on pre-defined relationships, or fully automated / 0 knowledge. Currently its a set of defined relationships applied in many combinations (try all and backtrack) which is not scalable.
Ideas:
Prolog comes to mind, but I am unsure of its scalability (100,000+ record samples) and I cant seem to find any documentation on interfacing with an external database (is encoding values as propositions the best/only way?)
Haskell or F# (don't know anything about these).
Thoughts? Recommendations?
Currently using:
PHP to interface with an Elasticsearch cluster. Most of the work is in the Elasticsearch queries with PHP being used for educated guesses and backtracking.
Clustering can either be based on pre-defined relationships, or fully automated / 0 knowledge. Currently its a set of defined relationships applied in many combinations (try all and backtrack) which is not scalable.
Ideas:
Prolog comes to mind, but I am unsure of its scalability (100,000+ record samples) and I cant seem to find any documentation on interfacing with an external database (is encoding values as propositions the best/only way?)
Haskell or F# (don't know anything about these).
Thoughts? Recommendations?