Analyzing Adverse Drug Events Using Data Mining Approach

23
MAJOR PROJECT ANALYZING ADVERSE DRUG EVENT USING DATA MINING APPROACH

Transcript of Analyzing Adverse Drug Events Using Data Mining Approach

Page 1: Analyzing Adverse Drug Events Using Data Mining Approach

MAJOR PROJECTANALYZING ADVERSE DRUG EVENT USING DATA MINING APPROACH

Page 2: Analyzing Adverse Drug Events Using Data Mining Approach

What is Adverse Drug event

????

- Adverse drug event is the result of adverse drug reaction.

- Adverse drug reaction is defined as any undesirable experience associated

with the use of a medical product in a individual.

- In short Adverse drug event is defined as an unwanted or unintended

reaction that results from the normal use of one or more medications

Page 3: Analyzing Adverse Drug Events Using Data Mining Approach

Why Adverse Drug Events

happens ???

The adverse drug event have started taking place more frequently and is

increasing because of several reasons like development of new medications,

increase in the use of medications for disease prevention, increased

coverage prescription medications.

Page 4: Analyzing Adverse Drug Events Using Data Mining Approach

How can ADEs be prevented

????

If a database of adverse drug events which took place anywhere in the

world is maintained and can be accessed by the authorities ,hospitals ,etc. ,

any possible risks to patient's life can be avoided as we are already aware of

the outcomes.

Page 5: Analyzing Adverse Drug Events Using Data Mining Approach

So , How can we help here

????

In this project, we plan on creating an Adverse drug reaction response system

which refers to the analytical tool which will be helping to analyze the

adverse drug reaction triples from the database.

Page 6: Analyzing Adverse Drug Events Using Data Mining Approach

What is to be done ??

Firstly the database is maintained using the already existing data or the data

provided by the user via. Interfaces for which we use the client-server

connection. Then, we use Apriori algorithm for mining the data which gives us

the potential drugs or drug pairs or triplets which can result into Adverse Drug

Events which initially starts with a transaction table.

Page 7: Analyzing Adverse Drug Events Using Data Mining Approach

Data saved and managed

using client-server

connection.

5 interfaces for entry but one database from which the data is mined.

Page 8: Analyzing Adverse Drug Events Using Data Mining Approach

Apriori Algorithm used for

mining data

The Apriori algorithm is a classic algorithm for learning association rules.

Association rule learning is a popular method for discovering relations

between variables in large databases.

Page 9: Analyzing Adverse Drug Events Using Data Mining Approach

Some definitions to

be aware of

Page 10: Analyzing Adverse Drug Events Using Data Mining Approach

What is the support ??

The support of an item (or set of items) is the number of transactions in which

that item (or items) occur.

Page 11: Analyzing Adverse Drug Events Using Data Mining Approach

What is the support threshold

??

The support threshold is defined by the user and is a number for which the

support for each item (or items) has to be equal or above for the support

threshold to be fulfilled.

Lets say in this example ,the value is 40 %.

Page 12: Analyzing Adverse Drug Events Using Data Mining Approach

What is a frequent item set

??

A frequent item set is an item set whose number of occurrences in the

transactions are above the support threshold.

Page 13: Analyzing Adverse Drug Events Using Data Mining Approach

Firstly, start with the

transaction database.

Transaction database contains the relationship status of medicines to cases

already present in the database.

Page 14: Analyzing Adverse Drug Events Using Data Mining Approach

From where we create a

candidate item set – Pass 1

We basically find the support values for each item i.e. drug in our case.

Page 15: Analyzing Adverse Drug Events Using Data Mining Approach

Now we create the frequent

item set – Pass 1

Where we keep the items , i.e. drugs whose support >= threshold support (40%

in this case).

Page 16: Analyzing Adverse Drug Events Using Data Mining Approach

Now we find the support

values for drug pairs – Pass 2

Here the item set contains all the possible permutations between the given

items and their support values are found.

Page 17: Analyzing Adverse Drug Events Using Data Mining Approach

Then again, we find the

frequent item set for Pass 2

Support >= threshold or min. support , else the item i.e. drug pair is

removed ; which in turn means that the removed pairs are not so

frequent.

Page 18: Analyzing Adverse Drug Events Using Data Mining Approach

Now moving on to drug

triplets – Pass 3

- Here, first we combine the elements to get triplets using the frequent item

set from pass 2, where the first element is same.

- Their support value is the min. [support{1st element in the combination from

pass 2), support (2nd element in the combination from pass 3)]

Page 19: Analyzing Adverse Drug Events Using Data Mining Approach

For frequent item set ,the sets

are pruned similarly – Pass 3

Support >= threshold or min. support

Page 20: Analyzing Adverse Drug Events Using Data Mining Approach

So in this way we find the frequent item set in

each pass and the result can be displayed after

each pass, in accordance to the user

requirements. So the user can see the risk in taking

the drugs as their effect and rest of the related

information is displayed.

Page 21: Analyzing Adverse Drug Events Using Data Mining Approach

Pseudo Code :-Pass 1

1.Generate the candidate itemsets in C1

2.Save the frequent itemsets in L1

3.Pass kGenerate the candidate itemsets in Ck from the frequent

itemsets in Lk-1

1. Join Lk-1 p with Lk-1q, as follows:

insert into Ck

select p.item1, p.item2, . . . , p.itemk-1, q.itemk-1

from Lk-1 p, Lk-1q

where p.item1 = q.item1, . . . p.itemk-2 = q.itemk-2, p.itemk-1 < q.itemk-1

2. Generate all (k-1)-subsets from the candidate itemsets in Ck

3. Prune all candidate itemsets from Ck where some (k-1)-subset of the candidate itemset is not in the frequent itemset Lk-1

4.Scan the transaction database to determine the support for each candidate itemset in Ck

5.Save the frequent itemsets in Lk

Page 22: Analyzing Adverse Drug Events Using Data Mining Approach

Pseudo Code (for our project) :-

In pass 1 , Candidate item sets are generated and then frequent item sets are

generated when the support values from candidate item sets are compared

to the specified min. support value which act as a threshold support value

and the items with support value less than this value is removed, which then

leads to pass 2 which uses the frequent item sets from pass 1 to create item

sets of entities pairs where only the elements with support value greater than

specified support value are kept giving us a new frequent item set . Then this

frequent item sets are used in pass 3 where the two elements with first

element same are combined to get triplets and the min. support value of the

taken two frequent set is taken and then compared to the min. specified

support value to get the frequent item sets for pass 3. So, here we can get the

output from all the passes depending upon the user requirements.

Page 23: Analyzing Adverse Drug Events Using Data Mining Approach

Supervisor – Ms. Megha Rathi

Submitted by – Rupal Bhardwaj (10503880)

Amisha Kathuria (10501832)

Batch – B11