Recommendation systems are now inherent for many business applications to take important business decisions. These systems are built based on the historical data that may be the sales data or customer feedback etc. Customer feedback is very important for any organization as it reflects the view, sentiment of the customers. Online systems allow customers to purchase products at a glance from any e-commerce website. Generally, the potential buyers check the review of the products to take informed decision of purchase. In this work, we attempt to build a recommendation model to find out the influence of a product on another product so that if a user purchases the influential product then the recommender system can recommend the influenced products to the users. In this paper, the recommendation system has been built based on association rule mining. We proposed a new association rule mining technique for quick decision-making and it gives better performance over Apriori algorithm which is one of the most popular approaches for association rule mining. The entire framework has been developed in Neo4j graph data model for doing the data modelling from raw text file and also to perform the analysis. We used real-life customer feedback data of amazon for experimental purpose.

https://link.springer.com/content/pdf/10.1007/s42979-021-00460-8.pdf

Recommendation of Influenced Products Using Association Rule Mining: Neo4j as a Case Study

Over the past decade’s several new concepts emerged to organize and query data over large Data Warehouse (DW) system with the same primary objective, that is, optimize processing speed. More recently, with the rise of BigData concept, storage cost lowered significantly, and performance (random accesses) increased, particularly with modern SSD disks. This paper introduces and tested a storage alternative which goes against current data normalization premises, where storage space is no longer a concern. By de-normalizing the entire data schema (transparent to the user) it is proposed a new concept system where query execution time must be entirely predictable, independently of its complexity, called, SINGLE. The proposed data model also allows easy partitioning and distributed processing to enable execution parallelism, boosting performance, as happens in MapReduce. TPC-H benchmark is used to evaluate storage space and query performance. Results show predictable performance when comparing with approaches based on a normalized relational schema, and MapReduce oriented.

SINGLE vs. MapReduce vs. Relational: Predicting Query Execution Time

Materialized views are heavily used to speed up the query response time of any data centric application. In the literature, the construction and dynamic maintenance of materialized views are carried out in a Binary Data Space where all attributes are given the same weight. Considering different weights may be particularly significant when similar queries are fired from multiple sites in a distributed environment, as taking into account the number of accesses to the different attribute values may reflect into the ability of tuning the materialized views accordingly. The methodology to construct weighted materialized view introduced in this paper is based on the association mining techniques, by applying it in a Non-Binary Data Space in distributed environments. The allocation of the views in the operating sites is also considered to a suitable use in distributed databases. Experimental results prove the superiority of proposed methodology on three bench mark datasets in terms of query Hit-Miss ratio and regulation of the view size with varying requirement of practical applications.

Construction and distribution of materialized views in Non-binary data space

Materialized views are heavily used to speed up the query response time of any data centric application. In literature, the construction and dynamic maintenance of materialized views are carried out in a Binary Data Space where all attributes are given the same weight. Considering different weights may be particularly significant when similar queries are posed by multiple users, as taking into account the number of accesses to the different attribute values may reflect into the ability of tuning the materialized views accordingly. The methodology to construct weighted materialized view introduced in this paper is based on the association mining techniques, by applying it in a Non-Binary Data Space. The proposed algorithm has been verified by simulation experiments with two benchmark datasets using practical transactional queries. The experimental results prove the superiority of our proposal in terms of query Hit-Miss ratio and flexibility of view size extendibility according to the requirement of practical applications.

Construction of Materialized Views in Non-Binary Data Space

View materialization using fuzzy MAX–MIN composition with association rule mining (VMFCA)

Analysis of data is an inherent part in the world of business to identify interesting patterns underlying in the data set. The size of the data is usually huge in the modern day application. Searching the data from the huge data set with a lesser time complexity is always a subject of interest. These data are mostly stored in tables based on relational model. Data are fetched from these tables using SQL queries. Query response time is an important quality factor for this type of system. Materialized view formation is the most common way of enhancing the query execution speed across industries. Different approaches have been applied over the time to generate materialized views. However few attempts have been made to construct materialized views with the help of Association based mining algorithms and none of those existing Association based methods measure the performance of the views in terms of both Hit-Miss ratio and view size scalability. This paper proposes an algorithm which generates a materialized view by considering the frequencies of the multiple attributes at a time taken from a database with the help of Apriori algorithm. Apriori algorithm is used to generate frequent attribute sets which are further considered for materialization. Moreover by varying the support count, changing the sizes of the frequent attributes sets; proposed methodology supports scalabilisalubrityty as well as flexibility. Experimental results are given to prove the enhanced results over existing inter-attribute analysis based materialized view formation.

Bibekananda Shit

Papers

Association Based Multi-attribute Analysis to Construct Materialized View

Construction and distribution of materialized views in Non-binary data space

Construction of Materialized Views in Non-Binary Data Space