Please use this identifier to cite or link to this item:
Title: Predicting Employability from User Personality using Ensemble Modelling
Authors: Sharma, Sahil
Supervisor: Garg, Deepak
Rana, Prashant
Keywords: Algorithms;CSED
Issue Date: 29-Jul-2015
Abstract: In field of machine learning numerous down to earth executions have been done for diverse purposes. Be that as it may, less work has been done in field of psychology research taking help of machine learning and data mining strategies. Big five identity model is the most broadly acknowledged psychology research identity display that is acknowledged in the field of applied psychology. Five variables can characterize one's identity as indicated by Big five model and they are Openness, Conscientiousness, Extrovertness, Agreeableness, and Neuroticism, subsequently the acronym OCEAN. In the data that has been prepared and tried for the outcomes that are given in this paper, aggregate of 19719 perceptions have been recorded where each individual has given their reactions to aggregate of 50 inquiries. 10 inquiries for every identity attribute were inquired. Theory has been taken to choose the order of individual, in the event that he/she is employable taking into account a few scopes of diverse identity components. At last, 6 machine learning models have been checked for their exhibitions to foresee the employability of 70-30 model. Models with 90% above precision have been checked with their changes and blends for checking which of the mix is anticipating the best result i.e. with most astounding exactness and ensemble model for this assignment is chosen. Group procedure verifies that the proposed model will perform on mixture of datasets with high precision. On the off chance that base models don't give adequate result, ensemble model will beat any base model with best result conceivable
Description: M.E. (Information Security-CSED)
Appears in Collections:Masters Theses@CSED

Files in This Item:
File Description SizeFormat 
3442.pdf1.05 MBAdobe PDFThumbnail

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.