Input: Feature_Vertex Object(FVO), Wordnet |
Output: The optimal action for FVO relationships; |
Procedure |
{ intialize ; |
/*implement -learning algorithm to get the action with the highest -value. /* |
∈ action_space = {similarity, contiguity, contrast, causality} |
/*Parallel.foreach used for parallel processing of all actions. /* |
Parallel.for(0; action_space.count(); => |
Get_Action_Reward([]); /*Function to implement (8). /* |
* * ( + γ * ()) |
}); // Parallel.For |
return Get_highest_-value_action(); /*select action with the maximum summed value of -values /* |
} |
/*execute (8) to get reward value. /* |
Get_Action_Reward( ) |
{ = (freq(.val)/FVO.Count()) * log(/); |
for (int = 0; < FVO.Count(); ++) |
{ = (freq(FVO[])/FVO.Count()) * log(/); |
SemRel() = Get_SemRel(.val, FVO[]) |
Sum += * SemRel(); |
} |
return + Sum; |
} |