Showing papers by "Parmeet Kaur published in 2018"

Graph Database for Recipe Recommendations

Abstract: Graph databases represent a paradigm shift from relational databases with a strong support for “ relationships”. As compared to relational databases which compute relationships at runtime, graph databases persist relationships for fast querying and data retrieval. This work presents a recipe recommender as a graph database, Neo4j application. Given any set of ingredients, this application recommends a variety of recipes with the help of a data set containing thousands of ingredients. Further based on availability of ingredients with a user, this application helps discover the list of possible dishes with these ingredients. In order to implement this application, ingredients and recipes have been crawled from cookery based websites using Python scripts. The crawled data has been inserted into the Neo4j database and subsequently inter-relationships between ingredients and recipes nodes have been analyzed. Execution of self designed queries has verified the time-efficiency of the proposed approach.

Empirical Investigation of Trends in NoSQL-Based Big-Data Solutions in the Last Decade

Abstract: The usage and popularity of NoSQL databases have sharply risen over the past decade due to their ability to handle a huge amount of data by employing scalable architecture, high availability and better performance than traditional relational database systems (RDBMS). In addition to reporting dynamics in NoSQL-database world, this paper focuses on presenting results from the perspective of developers. Stack Overflow provides a comprehensive technical niche with about 15 million technical questions, 8.1 million users and 25 million answers. In this paper, we aim to study variation in yearly trends of 20 NoSQL databases. To reveal the interest of the programmers we have investigated questions-asked and presented an unbiased Normal Interest Score by employing three parameters, first, the number of questions asked, second, mean views on a question and third, the mean score on a question. MongoDB, Cassandra, Redis, and Neo4j emerged as most popular databases in their respective families while NIS of all four of them is decreasing 2015 onwards. Additionally, we have also discussed how real-world events like publications, open-sourcing, mention in critical bills, version-release, acquiring ventures etc affect the interest corresponding to NoSQL databases over Stack Overflow. Results of this work will help database developers, database administrators for database selection, upgradation and maintenance.

Mobile Sink as Checkpoints for Fault Detection Towards Fault Tolerance in Wireless Sensor Networks

Abstract: A WSN consists of a large number of limited computation and storage capability wireless sensor nodes, which communicate wirelessly. These sensor nodes typical communicate in short range and collaborate to accomplish the network function. To increase the range of sensing and with the advent of MEMS, mobile sensors and sinks is the technology the world is moving to. This paper presents a network of mobile sensors and a sink. A mobile sink is selected as check-point to have the recoverability of the network. A Fuzzy Rule based system (FRS) is used to construct and select efficient static sensor nodes having adequate resources as Check Point Storage Nodes (CPSNs). The objective of FRS is to increase the probability of recovery of check-pointed data subsequent to a failure, thereby allowing a distributed application to complete its execution successfully. Simulations show FRS's better recovery probabilities in comparison to a random check-pointing arrangement.

An Empirical Comparison of MongoDB and Hive

Abstract: Most of the web and mobile applications today involve storage, processing and analysis of large datasets. The existing relational database systems are inadequate in handling the basic challenges introduced by these data-centric applications. This, consequently, has led to a new class of scalable and non-relational data management systems, referred to as NoSQL databases. NoSQL systems are characterized by their ability to scale horizontally and provide high availability. Apart from NoSQL, Hadoop framework and its constituent technologies are also synonymous with solutions for large data sets. This paper investigates the querying performance of a widely used NoSQL document store, MongoDB and compares its performance with respect to the Hadoop analytical language, Hive over a single node. The experimental results show that MongoDB yields a better performance than Hive for the considered dataset over a single node.

Empirical Evidence of Heuristic and Cost Based Query Optimizations in Relational Databases

Abstract: A SQL query may be expressed in different forms with varying execution costs; hence, it is essential to choose the best when performance is considered. There are many factors that affect the cost and performance of a query. For instance, a query on a relation without proper indexes can cause full relation scans, which in turn causes either performance or locking problems. Query asking for a relation held by another query can cause dead locks and an improperly formed query can also lead to greater complexity. In this context, the paper presents an empirical analysis of the heuristic and cost-based query optimization techniques. The execution cost of a number of queries was calculated and compared with execution plans obtained after optimization. In particular, we have experimentally checked the impact of query rewriting; the ordering of SQL joins and Select statements and replacement of joins with semi-joins. The experiments have yielded a clear insight into the effect of the mentioned optimizations.