What best describes weighted distributions in this context?

What best describes weighted distributions in this context?

• A. The percentage that one type will appear is based on the frequency of that type, divided by the total amount of tokens.
• B. The distribution assigns equal probability to all types.
• C. Weights are based on the word length.
• D. Weights correspond to syntactic roles.

Answer: (A)

Weighted distributions assign a probability to each type in proportion to how often that type occurs. In this context, the chance of a specific type appearing is its frequency divided by the total number of tokens. This creates an empirical distribution where P(type) = freq(type) / total_tokens, and all probabilities sum to 1. For example, if there are 120 tokens with 30 of one type, 50 of another, and 40 of a third, their probabilities would be 0.25, 0.4167, and 0.3333 respectively. This approach reflects the observed data rather than treating every type as equally likely or weighting by something like word length or syntactic role.

Weighted distributions assign a probability to each type in proportion to how often that type occurs. In this context, the chance of a specific type appearing is its frequency divided by the total number of tokens. This creates an empirical distribution where P(type) = freq(type) / total_tokens, and all probabilities sum to 1. For example, if there are 120 tokens with 30 of one type, 50 of another, and 40 of a third, their probabilities would be 0.25, 0.4167, and 0.3333 respectively. This approach reflects the observed data rather than treating every type as equally likely or weighting by something like word length or syntactic role.