Date of Award

Fall 2021

Publication Type

Thesis

Degree Name

M.Sc.

Department

Computer Science

First Advisor

C. Ezeife

Second Advisor

H. Fani

Third Advisor

M. Belalia

Keywords

Feature extraction, High utility sequential pattern Mining, Opinion mining, Sentiment matching

Rights

info:eu-repo/semantics/openAccess

Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.

Abstract

Extracting a group of features together instead of a single feature from the mined opinions, such as “{battery, camera, design} of a smartphone,” may yield higher profit to the manufactures and higher customer satisfaction, and these can be called High Profit Feature Groups (HPFG). The accuracy of Opinion-Feature Extraction can be improved if more complex sequential patterns of customer reviews are learned and included in the user-behavior analysis to obtain relevant frequent feature groups. Existing Opinion-Feature Extraction systems that use Data Mining techniques with some sequences include those referred to in this thesis as Rashid13OFExt, Rana18OFExt, and HPFG19_HU. Rashid13OFExt and Rana18OFExt systems use Sequential Pattern Mining, Association Rule Mining, and Class Sequential Rules to obtain frequent product features and opinion words from reviews. However, these systems do not discover the frequent high profit features considering utility values (internal and external) such as cost, profit, quantity, or other user preferences. HPFG19_HU system uses High Utility Itemset Mining and Aspect-Based Sentiment Analysis to extract High Utility Aspect groups based on feature-opinion sets. It works on transaction databases of itemsets formed using aspects by considering the high utility values (e.g., are more profitable to the seller?) from the extracted frequent patterns from a set of opinion sentences. However, the HPFG19_HU system does not consider the order of occurrences (sequences) of product features formed in customer opinion sentences that help distinguish similar users and identifying more relevant and related high profit product features.

This thesis proposes a system called High Profit Sequential Feature Group based on High Utility Sequences (HPSFG_HUS), which is an extension to the HPFG19_HU system. The proposed system combines Feature-Based Opinion Mining and High Utility Sequential Pattern Mining to extract High Profit Feature Groups from product reviews. The input to the proposed system is the product reviews corpus. The output is the High Profit Sequential Feature Groups in sequence databases that identify sequential patterns in the features extracted from opinions by considering the order of occurrences of features in the review. This method improves on existing system's accuracy in extracting relevant frequent feature groups. The results on retailer’s graphs of extracted High Profit Sequential Feature Groups show that the proposed HPSFG_HUS system provides more accurate high feature groups, sales profit, and user satisfaction. Experimental results evaluating execution time, accuracy, precision, and comparison show higher revenue than the tested existing systems.

Share

COinS