Big Data has gained attention due to its significant impact on society. With the growth of IoT, data has grown exponentially. Big Data has played an essential role in decision-making. After collecting a large volume of data, one of the critical tasks is performing analysis. As a result, there is a rise in the adoption of data visualization, where the raw data is converted into useful information through research and the generation of patterns. The representation of the natural numbers offers an easier way to interpret the collected data, and thus, it becomes easier to comprehend the significance of the data collected. While data visualization is significant, it faces numerous challenges; therefore, this paper aims to provide the different challenges of Big Data visualization.

IoT generates a large volume of data in different formats. Exponential data generation has resulted in challenges in capturing, managing, storing, analyzing, and visualizing data. This necessitates the implementation of exemplary database architecture to store data. The paper aims to evaluate the various database architectures.

There is a surge in the usage of IoT platforms, which has increased the number of interconnected devices worldwide. There are different challenges faced with adopting the IoT platform, and this paper will evaluate the storage challenges.

A polystore system is composed of multiple heterogeneous databases which are used to store data that best suites its features for faster processing and querying. A single database does not fit all data formats, which has resulted in the development of a polystore system that leverages databases belonging to different categories such as relational, NoSQL, and NewSQL. Retrieving data from relational databases for analytics is a costly operation. This paper presents an extension to a polystore system which queries data from different relational databases using a standard in-memory data format for analytical queries with low latency.

Big Data analytics is used in decision making. It involves heavy computation to obtain exact answers. To alleviate this problem, approximate query processing (AQP) was adopted, which provides approximate results with error bounds. The AQP models which have been proposed are supported only by a single database. In an organization, big data is stored in multiple databases that have different data models. This research aims to provide AQP as a middleware solution using query optimization for heterogeneous databases.

As the world continues to encounter challenges related to public health and economic well-being, technology and big data can be used to carry out analyses and provide predictions concerning the progression of the COVID-19 pandemic. This information allows virologists and epidemiologists to understand the Coronavirus better. Historically, medical experts relied on hospital records and medical publications to manage pandemics. However, this approach proved to be slow and inefficient when it comes to predicting a disease's course. Using contemporary technologies for collecting and analyzing information is a fast and efficient way.

As the number of databases continues to grow data scientists need to use data from different sources to run machine learning algorithms for analysis. Data science results depend upon the quality of data been extracted. The objective of this research paper is to implement a federated query processing framework which extracts data from different data sources and stores the result datasets in a common in-memory data format. This helps data scientists to perform their analysis and execute machine learning algorithms using different data engines without having to convert the data into their native data format and improve the performance.

Similarity and dissimilarity measures play an important role in pattern classification and clustering. For a century, researchers have searched for a good measure. Here, we review, categorize, and evaluate various binary vector similarity/dissimilarity measures for character recognition. One of the most contentious disputes in the similarity measure selection problem is whether the measure includes or excludes negative matches. While inner-product based similarity measures consider only positive matches, other conventional measures credit both positive and negative matches equally. Hence, we propose an enhanced similarity measure that gives variable credits and show that it is superior to conventional measures in an offline handwritten character recognition application. Finally, the proposed similarity measure can be further boosted by applying weights and we demonstrate that it outperforms the weighted Hamming distance.

Handwriting originates from a particular copybook style such as Palmer or Zaner-Bloser that one learns in childhood. Since questioned document examination plays an important investigative and forensic role in many types of crime, it is important to develop a system that helps objectively identify a questioned document's handwriting style. We proposed a handwriting analysis system that can assist a document examiner in the identification of the writer's handwriting style and therefore of his/her origin or nationality. We collected 33 English alphabet copybook styles from 18 countries. Here, we extend the analysis using several data mining techniques to discover important information that can be gleaned from a handwriting copybook style image database, e.g., the most information bearing alphabet characters for the purpose of copybook style identification and the relationship between geographical regions and similarity based clusters of copybook styles.

Classifying an unknown input is a fundamental problem in pattern recognition. A common method is to define a distance metric between patterns and find the most similar pattern in the reference set. When patterns are in binary feature vector form, there have been two approaches to improve the performance over the equal-weighted Hamming distance metric. One is to give different weights to different features using an optimization technique, and the other is to use a similarity measure that gives full credit to features present in both patterns and the less credit to those absent from both patterns. Both approaches have been reported to perform better than the na ve Hamming distance approach. In this paper, we propose to combine these two approaches using a genetic algorithm to optimize weights. Experimental results show that this method is superior to conventional measures in an OCR application.