A dataset of length n can be divided *log n* times until everything is completely divided. Therefore, the search complexity of binary search is O(log n).

A binary search can be performed in an iterative approach. Unlike calling a function within the function in a recursion, this approach uses a loop.

function binSearchIterative(target, array, left, right) {while(left < right) {let mid = (right + left) / 2;if (target < array[mid]) {right = mid;} else if (target > array[mid]) {left = mid;} else {return mid;}}return -1;}

In a recursive binary search, there are two cases for which that is no longer recursive. One case is when the middle is equal to the target. The other case is when the search value is absent in the list.

```
binary_search(sorted_list, left_pointer, right_pointer, target)
if (left_pointer >= right_pointer)
base case 1
mid_val and mid_idx defined here
if (mid_val == target)
base case 2
if (mid_val > target)
recursive call with left pointer
if (mid_val < target)
recursive call with right pointer
```

In a recursive binary search, if the value has not been found then the recursion must continue on the list by updating the left and right pointers after comparing the target value to the middle value.

If the target is less than the middle value, you know the target has to be somewhere on the left, so, the right pointer must be updated with the middle index. The left pointer will remain the same. Otherwise, the left pointer must be updated with the middle index while the right pointer remains the same. The given code block is a part of a function `binarySearchRecursive()`

.

function binarySearchRecursive(array, first, last, target) {let middle = (first + last) / 2;// Base case implementation will be in here.if (target < array[middle]) {return binarySearchRecursive(array, first, middle, target);} else {return binarySearchRecursive(array, middle, last, target);}}

Binary search performs the search for the target within a sorted array. Hence, to run a binary search on a dataset, it must be sorted prior to performing it.

The binary search starts the process by comparing the middle element of a sorted dataset with the target value for a match. If the middle element is equal to the target value, then the algorithm is complete. Otherwise, the half in which the target cannot logically exist is eliminated and the search continues on the remaining half in the same manner.

The decision of discarding one half is achievable since the dataset is sorted.

The binary search algorithm takes time to complete, indicated by its `time complexity`

. The worst-case time complexity is `O(log N)`

. This means that as the number of values in a dataset increases, the performance time of the algorithm (the number of comparisons) increases as a function of the base-2 logarithm of the number of values.

Example: Binary searching a list of 64 elements takes at MOST `log2(64)`

= 6 comparisons to complete.

The binary search algorithm efficiently finds a goal element in a sorted dataset. The algorithm repeatedly compares the goal with the value in the middle of a subset of the dataset. The process begins with the whole dataset; if the goal is smaller than the middle element, the algorithm repeats the process on the smaller (left) half of the dataset. If the goal is larger than the middle element, the algorithm repeats the process on the larger (right) half of the dataset. This continues until the goal is reached or there are no more values.

Linear search can be used to search for the smallest or largest value in an unsorted list rather than searching for a match. It can do so by keeping track of the largest (or smallest) value and updating as necessary as the algorithm iterates through the dataset.

```
Create a variable called max_value_index
Set max_value_index to the index of the first element of the search list
For each element in the search list
if element is greater than the element at max_value_index
Set max_value_index equal to the index of the element
return max_value_index
```

For a list that contains `n`

items, the best case for a linear search is when the target value is equal to the first element of the list. In such cases, only one comparison is needed. Therefore, the best case performance is O(1).

Linear search runs in linear time and makes a maximum of `n`

comparisons, where `n`

is the length of the list. Hence, the computational complexity for linear search is O(N).

The running time increases, at most, linearly with the size of the items present in the list.

A linear search can be expressed as a function that compares each item of the passed dataset with the target value until a match is found.

The given pseudocode block demonstrates a function that performs a linear search. The relevant index is returned if the target is found and -1 with a message that a value is not found if it is not.

```
For each element in the array
if element equal target value then
return its index
if element is not found, return
“Value Not Found” message
```

A function that performs a linear search can return a message of success and the index of the matched value if the search can successfully match the target with an element of the dataset. In the event of a failure, a message as well as `-1`

is returned as well.

```
For each element in the array
if element equal target value then
print success message
return its index
if element is not found
print Value not found message
return -1
```

A linear search can be modified so that all instances in which the target is found are returned. This change can be made by not ‘breaking’ when a match is found.

```
For each element in the searchList
if element equal target value then
Add its index to a list of occurrences
if the list of occurrences is empty
raise ValueError
otherwise
return the list occurrences
```

*Linear search* sequentially checks each element of a given list for the target value until a match is found. If no match is found, a linear search would perform the search on all of the items in the list.

For instance, if there are `n`

number of items in a list, and the target value resides in the `n-5`

th position, a linear search will check `n-5`

items total.

Despite being a very simple search algorithm, linear search can be used as a subroutine for many complex searching problems. Hence, it is convenient to implement linear search as a function so that it can be reused.

The best-case performance for the Linear Search algorithm is when the search item appears at the beginning of the list and is O(1). The worst-case performance is when the search item appears at the end of the list or not at all. This would require N comparisons, hence, the worse case is O(N).

The Linear Search Algorithm performance runtime varies according to the item being searched. On average, this algorithm has a Big-O runtime of O(N), even though the average number of comparisons for a search that runs only halfway through the list is N/2.

The Linear Search algorithm has a Big-O (worst case) runtime of O(N). This means that as the input size increases, the speed of the performance decreases linearly. This makes the algorithm not efficient to use for large data inputs.