Today we continue our discussion on natural language processing with a particular word ontology called FrameNet. The idea behind FrameNet is that the definition of a word is incomplete without the relations to conceptual structures and patterns of beliefs, practices, institutions, images etc. which are referred to as semantic frames. - courtesy Filmore 2003. Therefore meanings are relative to Frames. The scope of FrameNet includes Events: for example related to birth or death as noun, Relations: for example being relevant as in adjective or in a personal relationship as in a noun, States: for example on or off as in preposition, and Entities: for example a device. Therefore the words are related to notions that are represented by frames and the meaning is not complete without these relationships. A Semantic Frame is the description of an event, relation, entity or state and its participants thereby capturing the essential knowledge of a given word sense. Roles within the frame are called Frame Elements. Words that evoke the frame are called Lexical units. For example cooking appears in a frame called 'Apply_heat' where the frame elements are the cook, food, container and heating_instrument. and words that evoke it such as fry, bake, boil and broil are lexical units.
FrameNet comprises of over 1195 Semantic frames with approximately 12 frame elements per frame. It includes12,989 lexical units which represent connection between one lemma + POS and one frame. It includes 198,932 manual annotations of corpus examples and about 1774 frame to frame relations. Since the relations involve multiple inheritance, it forms a lattice. It is also constructed bottom-up. It comes with semantic type labels such as animate, human, positive_judgement and assumes construction grammar which describe the grammatical characteristics and semantic import of each construction ( Filmore, Goldman, Rhodes ). It also includes metaphor and metonymy.
FrameNet is not a formal ontology. It includes non-lexical frames that have been made up as needed to connect the nodes in certain cases. It attempts to include cross linguistic differences such as the difference between commercial_transaction versus a criminal_process although both describe methods.
Why is it an improvement over semantic ontologies ?
1) Frame is a useful level of abstraction
2) Roles are more meaningful than semantic labels
3) Frames represent conceptual gestalts - more than just the sum of their parts.

#codingexercise
Insert into a circular linked list that is sorted but where we are given a random node from the list not the head.

Void insertCircular(ref node start, node target)

{
var head = FindHead(ref node start, node target);
If (head == null) {head = target; target.next = target; return;}
Node prev = null;
Node curr = head;
while(curr.data < target.data && curr.next != head){
    Prev = curr;
    Curr = curr.next;
    }
 If (prev == null){
    If (curr.data < target.data){
         InsertAt(cur, target);
    }else{
         InsertAtStart(head, target);
     }
 }else{
   if (curr.data < target.data) {
        InsertAt(cur, target);
   }else{
        InsertAt(prev, target);
   }
 }
}

Node FindHead(ref node start, node target)
{
Node head = null;
if (start == null) return head;
If (start.next == start) return start;
Node curr = start;
min = curr.data;
head = start;
while(curr.next != start){
    if (curr.data < min)
    {
         head = curr;
         min = curr.data;
    }
    curr = curr.next;
}   
 if (curr.data < min)
    {
         head = curr;
         min = curr.data;
    }
return head;
}
Alternatively, we could also insert by iterating the circular list once to find the node that is closest to the given node in its value.