Abstract: Provided are techniques for identifying and deleting unnecessary code lines. A plurality of global behavioral rules are stored, where each of the global behavioral rules includes a behavioral pattern code fragment, a plurality of recommendations, and a recommendation score for each of the plurality of recommendations. It is determined that a code fragment in a portion of code matches a behavioral pattern code fragment of a global behavioral rule of the plurality of global behavioral rules. A recommendation from the plurality of recommendations of the global behavioral rule is provided based on the recommendation score of that recommendation. In response to the recommendation being applied, a recommendation score of the recommendation is increased. In response to the recommendation being rejected, an action taken is identified, and the action taken is stored as a new recommendation for the behavioral pattern code fragment of the global behavioral rule.

Abstract: Provided are techniques for identifying and deleting unnecessary code lines. A plurality of global behavioral rules are stored, where each of the global behavioral rules includes a behavioral pattern code fragment, a plurality of recommendations, and a recommendation score for each of the plurality of recommendations. It is determined that a code fragment in a portion of code matches a behavioral pattern code fragment of a global behavioral rule of the plurality of global behavioral rules. A recommendation from the plurality of recommendations of the global behavioral rule is provided based on the recommendation score of that recommendation. In response to the recommendation being applied, a recommendation score of the recommendation is increased. In response to the recommendation being rejected, an action taken is identified, and the action taken is stored as a new recommendation for the behavioral pattern code fragment of the global behavioral rule.