Assume that a data file has an index consisting of $\text{N}$ items, where $\text{N}$ is large. If a binary search of the index is used to find an item, then, of the following, which best approximates the mean number of comparisons required to locate a specific index entry?
- $\text{(N}+1) / 2$
- $\text{N(N}-1) / 2$
- $\left(\log _2 \text{N}\right)-1$
- $\text{N} \log _2 \text{N}$