Credibility Evaluation of User-generated Content using Novel Multinomial Classification Technique

Abstract

Awareness about the features of the internet, easy access to data using mobile, and affordable data facilities have caused a lot of traffic on the internet. Digitization came with a lot of opportunities and challenges as well. One of the important advantages of digitization is paperless transactions, and transparency in payment, while data privacy, fake news, and cyber-attacks are the evolving challenges. The extensive use of social media networks and e-commerce websites has caused a lot of user-generated information, misinformation, and disinformation on the Internet. The quality of information depends upon various stages (of information) like generation of information, medium of propagation, and consumption of information. Content being user-generated, information needs a quality assessment before consumption. The loss of information is also necessary to be examined by applying the machine learning approach as the volume of content is extremely huge. This research work focuses on novel multinomial classification (based on multinoulli distribution) techniques to determine the quality of the information in the given content. To evaluate the information content a single algorithm with some processing is not sufficient and various approaches are necessary to evaluate the quality of content. We propose a novel approach to calculate the bias, for which the Machine Learning model will be fitted appropriately to classify the content correctly. As an empirical study, rotten tomatoes’ movie review data set is used to apply the classification techniques. The accuracy of the system is evaluated using the ROC curve, confusion matrix, and MAP.

Web Service Implementation “AI Yoga Trainer Using Human Pose Estimation” for Web Discovery

“…The importance of maintaining proper posture for general health and well-being has increased with the prevalence of sedentary lifestyles [3] [4]. Poor posture can cause several muscular-skeletal issues and have a detrimental influence on physical performance numerous muscular-skeletal problems [5] [6], such as back discomfort, restricted mobility, and poor physical performance, can be brought on by poor posture [7] [8]. Keeping appropriate posture in yoga as well as in daily life has become increasingly important to fight these difficulties [2].…”

Performance analysis of Machine Learning Algorithms to classify Software Requirements

“…In this research paper, we aim to classify functional and non-functional software requirements using different algorithms: Support Vector Machine (SVM), Logistic Regression (LR), Decision Tree (DT), Random Forest (RF), and Multilayer Perceptron (MLP) Neural Networks to test how classification algorithms perform and how it can improve [10] [11]. We present an effective methodology for pre-processing the text data and applying machine learning algorithms to classify the requirements [12] [13]. We compare the performance of the algorithms and provide insights into the accuracy and limitations of each algorithm [14] [15].…”

Novel Approach to Design a Model for Software Effort Estimation Using Linear Regression

“…After that, it examines previous studies on various methods for estimating efforts, such as parametric methods, heuristic methods, and model-based methods [3] [4]. Lastly, it considers recent developments and difficulties in effort estimating and makes recommendations for future study lines [5] [6]. The ramifications for software developers who are thinking about how to precisely estimate project tasks are covered in the paper's conclusion [7] [8].…”