Saudi Electronic University Machine Learning Techniques Neural Network Report

User Generated

nmvm12345

Computer Science

Saudi electronic university

Description

this file assignment 1

find in the attached file

the seem data set continue assignment

i want continue the seem data set larg face book in assignment 1

Project: Assignment 2 Lecture Notes for CSC 563 7 Implementation with Machine Learning Techniques

 Write a report and present it.

The report should include: 

Introduction 

Method 

Experiment and results 

Discussion

Unformatted Attachment Preview

Name:Abdulaziz hazazi id:441106032 assignment 1 Link Prediction Dataset: Facebook Large page-page network Introduction Facebook is a social media platform that has lots of registered users. These users in turn connect to other users through friend requests and join a number of pages, whether it is for fun, politics, among others. This creates a social network that involves interaction between different organizations, political parties, organizations and people. These interactions create a lot of data which can be represented as graphs, where the entities represent the nodes and the interaction or relationships represent the links. Description of problem and objectives Having all these data in Facebook from these different entities, would you be curious to know which pair of entities or nodes will form a relationship or link in future? These relationships are described in terms of mutual likes. Now that presents our problem and this is a classic example of what link Prediction can do. The problem here is trying to identify unconnected entities that will form relationships in the future. Link Prediction serves to understand how entities relate and probably what these entities would do jointly. For example, using link Prediction, we can predict which people might attend a political rally, or attend an entertainment event together. We can even predict if two unconnected friends will become friends. And how does this work? Since we have connected pairs and nodes, we will turn this problem into a machine learning problem then use it to do predictions on unconnected nodes. Database selection and description We have a graph dataset downloaded from https://archive.ics.uci.edu/ml/datasets/Facebook+Large+Page-Page+Network# That contains data from four categories, politicians, governmental organizations, television shows and companies. These datasets have the nodes and the target variables as well. Objective: To build a Link Prediction that would predict future links between these Facebook pages. The nodes are shown below; Fig 1.0: A python snippet showing relationships between different nodes In fig 1.0 above, the id_1 and id_2 represents the nodes that will form the graph. Nodes in one row of the data frame represent links. Each node represents a Facebook page as shown below Fig 1.2: A python snippet showing the nodes and their target names In fig 1.2 above, node 22329, which represents NHTSA, which is in government category, is connected to City of Los Altos Police Department. This is one link. These page types will also be connected to other pages to forma network as the one shown below to form a very dense networks to form a graph. Fig 1.3: The graph network of Facebook large-large dataset The database selected has a number of nodes. These nodes which are blue represent the Facebook pages, while the black lines represent the links or the relationships between the pages. From fig 1.3 we can see the connected nodes. We shall then retrieve the connected pairs and connected pairs from the graph. The connected pairs represent the entities or pages that are connected while the unconnected pairs represent the entities that are not connected. The connected pairs will have a lot of information. It is from these connected pairs that we will mine information that will help us create a model that will help us predict if there will be links or likes between the unconnected pairs. From fig 1.3, we can see that there are lots of unconnected nodes at time T. These nodes may or may not be connected, but the model will tell us if they will be connected at time (T+t). The connected pairs will create predictor variables. These connected pairs will further be split into training and test datasets just as any other machine learning problem. We shall then retrieve the unconnected nodes using a variety of techniques like adjacency matrix. We shall then use an algorithm like in logistic regression or random forests to train the model using the training samples retrieved from graph validate it using the validation data. If the accuracy is acceptable, like more than 90% then we can use it to predict links between two unconnected nodes, which in this case represent the Facebook pages from the four categories introduced above.
Purchase answer to see full attachment
User generated content is uploaded by users for the purposes of learning and should be used following Studypool's honor code & terms of service.

Explanation & Answer

Kindly see attached

Name:Abdulaziz hazazi

id:441106032

assignment 2

Machine Learning techniques
Dataset: Facebook Large page-page network

Introduction
Nowadays, people used social media for various purposes. One of the most used social
media was Facebook application. Facebook delivered useful links and entities that users enjoyed
in different ways such us...


Anonymous
Really useful study material!

Studypool
4.7
Trustpilot
4.5
Sitejabber
4.4

Similar Content

Related Tags