Given the scenario where a company logs data from its application in large files and uses these logs for regular analytics for three months after generation, and then accesses these logs infrequently for up to a year, which persistence strategy would be most effective?

$1.$

Utilize a graph database for the entire duration to maximize the efficiency of connected data queries.

$2.$

Store the logs in a high$-$performance NoSQL database for three months, then transfer them to a cheaper, archival$-$oriented storage solution for the infrequent access period.

$3.$

Use a relational database for both the initial three months of active analytics and the subsequent year of infrequent access.

$4.$

Maintain all data in an in$-$memory database for fast access during both the initial three months and the subsequent infrequent access phase.