-
Notifications
You must be signed in to change notification settings - Fork 1
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
fsw similarity metrics #3
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Left some comments, but most importantly:
Please run closest_matches
to show what this metric actually does.
There are many unjustified numbers in here, but the results could speak for themselves
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
almost perfect. minor corrections
self.weight["positional"] * dis.euclidean(hyp["position"], ref["position"])) | ||
hyp_class = next((i for i, r in enumerate(self.symbol_classes.values()) if shape1 in r), None) | ||
ref_class = next((i for i, r in enumerate(self.symbol_classes.values()) if shape2 in r), None) | ||
distance = distance + abs(ref_class - hyp_class) * self.weight["class_penalty"] |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
i feel like this might be too lenient - in my opinion, same class = 1, different class = 0
@@ -14,45 +14,45 @@ def __init__(self): | |||
super().__init__("SymbolsDistances") | |||
self.symbol_classes = { | |||
'hand_shapes': range(0x100, 0x205), | |||
'contact_symbols': range(0x205, 0x221), | |||
'movement_paths': range(0x221, 0x2FF), |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
why did you remove all of them? even if removed, maybe leave them as commented out.
hyp_veq = tuple(val * weight for val, weight in zip(hyp_veq, [self.weight["shape"], self.weight["angle"], | ||
self.weight["facing"], self.weight["parallel"], | ||
self.weight["positional"]])) | ||
ref_veq = tuple(val * weight for val, weight in zip(ref_veq, [self.weight["shape"], self.weight["angle"], |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
this could also be a function btw weigh_vector
closes #1