Sunteți pe pagina 1din 4

Ma.

Nikkie Evalla

Mr. Ezra Anglo

BSIT2-6

Data Structure

Reflection Paper on Data structures: Binary Search Tree

In the following paper I will discuss what I have learned from watching an informative video by
mycodeschool entitled Data structures: Binary Search Tree. The purpose of this reflection paper and of the source
video as to how I sees it is to explain the Binary Search Tree. I decided to include a comparison of an array, linked list
and binary search tree running time for certain operations which is of course I also learnt from the said video to
further explain the topic.
As stated from our source, Binary search tree is a special kind of binary tree that organizes data for quick
search and update. This is just an introduction much like giving meaning to our main topic. This is followed by the
overview of two well-known data structure used for storing modifiable collection namely; an array and linked list. Use
an array or linked list as a modifiable collection to be able to search, insert and remove an element
Before anything else the source had a comparison amongst the two well-known data structure together with
the binary search tree. We will compute the running time for these operations; searching, inserting, and removal of
records of the two.
Starting off with the array, how integers will be stored in it is shown. This is done by creating a large enough
array to store the records marking the end of the list to determine if an index still contains a record and so as to say
that those after the mark are available space or empty indexes. For example, if we have index 18 marked as the end
then that means we have 19 records (of course 0 is included). To search let say variable x in the records we have to
scan the array from index 0 up to the marked end, if what we are searching for unfortunately is in the end as to what I
understood from the video is that we need to look through all the elements in the array. So let n be the number of
elements we have in our array therefore the running time of an array for the operation of searching will be
proportional to the number of elements n in it. The larger the array the higher the time it will consume to search for a
given x. The source denoted this as Search()= o()
Our previously done marker end will be very vital in the process of inserting a new record to the array; as
stated above those index that follows the maker end are all available space. This is where we will be adding the new
record, this is done by incrementing the marker end. After incrementing we can now add the new record which is an
integer in the case shown in the source video. For this operation (which is inserting) time taken will not depend on the
number of elements instead it is constant. It can be denoted as Insert()= o(1)

The next operation deleting is quite a crucial one. Say for example we want to remove the record in index 4,
in order to do that well have to shift all records to the right of index 4 one index to the left. Lastly we will decrement
our marker end thus resulting to a running operation of Remove()= o().
Moving forward to linked list this data structure runs almost the same as an array therefore the running time
of the operation search is also Search()= o() but in this case we are pertaining to number of nodes in the linked
list. In order to search we will have to traverse all the records of the whole list same as to an array we may need to
look through all the nodes in the linked list. While the cost of insertion in a linked list has two process either
Insertion()= o(1) if its at the head or Insertion()= o() at the tail. It is advised to insert at the head to keep the
running cost low and constant. In the process of removal we may have to traverse the linked list and search the
record checking at all the nodes Remove()= o().
We may think of these two as a suitable data structure for searching, inserting and removing. Even me I
would think of using it but after watching the first 5 minutes of the source video I doubt if you still have your initial
opinion regarding this. It has been clearly stated that in reality any of the two is not a practical method to use
although insertion between them is fast. One point that we may not have noticed is that these may consume a
relatively large amount of time if we have large collection of records. How about a record for a barangay with 1000
families most of them consisting of 4 members gaining a total of 4000 individual records? If we will go for an
estimated 1 second per search then we will end up consuming 4000 seconds searching for that certain x.
Computer nowadays deal with a vast amount of data. I couldnt just imagine how awful it was for a computer
to use either of the two with that number of data. The source video presented binary search as a solution for this
problem yet there are things needed to be considered when using this. First is that what you will be using shall be an
array and that your array should be sorted resulting at BinarySearch()= o(log). I have mentioned that the source
uses a simple integer as its data type so sorting would be easy except when using string or any other data type we
should be able to binary search it based on some property or some key of the records. Although running cost of this
will be minimized especially when using a sorted array keeping it sorted is a hard task cause every time you will carry
out insertion or removal you have to make sure that it is sorted afterwards. Searching the position to where a certain
say x for example shall be put is very easy and fast, costing only o(log) but despite this insertion and removal
requires shifting of all the records from the index where well be inserting or removing a record. This shifting of
records will cost o() thus Insert()= o() and Remove()= o(). Using binary search for comparing and searching is
very practical yet the cost of insertion and removal is still unacceptable. The source video seeks for another practical
method to use with an acceptable running time for all the stated operations.
So summarizing everything we are looking for the most practical means of handling such
operations, this is exactly where the Binary Search Tree comes in. Binary search tree (BST) offers the cost of these

three operations in lowest value possible that is o(log) for average cases and will hit as low as o() for its worst
cases. Of course we can as much as possible avoid worst case scenarios by making sure that the tree is always
balanced.
Moving on lets tackle first how cost of binary search tree operations are minimized. By definition, binary
search tree is a binary tree in which for each node, value of all the nodes in left sub tree is lesser or equal (to handle
duplicates) and value of all the nodes in right sub tree is greater. In binary tree each node can have at most two
children. The source video illustrated binary search tree as a circle on top branching out to two more circles (one on
the left and one on the right) which in return also branches out into two more circles. The illustration has the data type
of integer with number 15 placed on the topmost circle. To check if the illustration is correct we divide the tree into
two; left and right after that we check if all the values of the circle on the left side is lower or equals the value on the
topmost of the tree (in our case its 15). The values on the right side is as follows; 10, 8 and 12 these are less than 15
so the left side of the tree is correct. Next is the same as the left side but this time we have to make sure that the
values in right side are always greater than 15. The illustration contains 20, 17 and 25 obviously these are greater
than 15 so the right side of the tree is also correct. By the way the source video states that the topmost value (15 in
our illustration) is called the root.
How is it really possible for binary search tree to have a cost of running as o(log) for searching, inserting
and deleting? First lets discuss search, this operates by comparing what you are searching for with the root. Lets
use the previous illustration for this. I would like for us to try and search for 12 so comparing 12 to 15 it is lesser in
this case we will continue searching for 12 on the left sub tree since weve known that left sub tree contain values
lesser than the root. The advantage of this is that we can reduce the elements we need to search through as we are
moving to the next steps making it a lot faster. The next root is 10 so do the same steps again, 12 is greater than 10
proceed the search to right sub tree we have root 12 in there and its equal to what we are searching for so the search
ends. We can experience the worst case if the tree is unbalanced hence reducing only one elements from the search
space. It only works as if it is somewhat a linked list so we experience much slower searching. The same goes for
inserting a new record to the binary tree, we start off comparing the root to the value we will add. If it is greater then
proceed to right else proceed to left sub tree. If there is still a child continue comparing until you reached where it
should be added place it as a child either to the left or to the right. Lastly for the deletion it is much like searching with
just a little bit of adjustment of nodes or links.
Clearing up the issues we have encountered there are 3 data structures that can be used for searching,
inserting and deleting records; an array, linked list and binary search tree. It is advisable to use binary search tree
among them because it has lower cost of running operations. Binary search tree encounters worst case wherein the
cost of running operations rose to a higher level and to avoid this we should make sure that the tree is always

balanced. Inserting and deleting records causes binary search tree to be unbalanced all we have to do is to restore
the balancing.

S-ar putea să vă placă și