15 research outputs found
Yelp analytics
No full textYelp is a website and mobile app which publishes crowd-sourced reviews about local businesses. In this thesis, we analyze data about restaurants from Yelp, specifically the reviews, to predict the star-ratings of the restaurants based on the contents of the reviews. Our results are based on performing sentiment analysis on the reviews, which involves determining whether a review is positive or negative. Various machine learning techniques were applied to the data after appropriate extraction of linguistic features, to create classification models, and to predict star---ratings based on these models.M.S.Includes bibliographical referencesby Aayush Agrawa
Author identification system
No full textAbstract: Every one of us has different approach to speak and write, and there exists a long history of linguistic and stylistic analysis into authorship attribution. In last year’s, practical application for author identification have grown in area such as computer forensic(linking intercepted message to each other and to find rebel), criminal law(identifying author of payoff notes and harassing letter), civil law and computer security (tracking author of computer source code). This paper proposes the implementation of author identification system. This proposed system is based upon the principles and concepts of text analysis. For ensuring maximum accuracy in identifying author of the document we will be using TF-IDF algorithm which consists of extraction of features from the text, scoring these features and comparing them with a set of scores stored in the corpus
Using Sign-Language as an Input Modality for Microtask Crowdsourcing
No full textSeveral input types have been developed in different technological landscapes like crowdsourcing and conversational agents. However, sign language remains one of the input types that has not been looked upon. Although numerous amount of people around the world use sign language as their primary language, there have not been many efforts to include them in these technological landscapes. In this thesis, we hope to draw attention to and take a step towards the inclusion of deaf and mute people in microtask crowdsourcing. We identify some of the existing technical and research gaps in the current architectures for Sign Language Recognition/Translation in a real-time setting. Next, we determine various microtasks which can be adapted to use sign language as input, keeping in mind the challenges it introduces. We, then, investigate the effectiveness of a system that uses sign language as input by building a web application - SignUpCrowd - for microtask crowdsourcing, namely Visual Question Answering and Tweet Sentiment Analysis tasks, and comparing it with already prevalent input types such as text and click. This comparison with different popular input types will help understand how much of a difference there is for sign language as input. In addition, it will also show the preference of input types for the particular microtasks. For this, we developed three web applications with different input types and conducted a between-subject experimental study on Prolific wherein a number of workers (N=240) were asked to perform the above-mentioned tasks using sign language, text, and click input. Our results indicate that, in terms of task completion time and task accuracy, sign language as an input modality in microtask crowdsourcing is not significantly different from other, commonly used, input types. We also noticed that people's input type preference for the given microtasks for sign language was more than text input. Although people with no knowledge of sign language found it difficult, this input modality aims at a different target audience. This shows us that there is scope for sign language as an input type for microtask crowdsourcing among people, and paves the way for more efforts for the introduction of sign language in real-world applications.https://osf.io/n8pca/?view_only=fc7bf6ab55d6482f83ff2729c25b937fComputer Scienc
Stochastic maze solving under the geometric amoebot model
No full textIn this work we give:
a fully-asynchronous, local and distributed stochastic algorithm for collective maze solving under the Geometric Amoebot model of self-organizing particle systems; and
an algorithm to verify phototaxing in arbitrarily large systems given the set of possible configurations are known.This work was accepted to the annual Graduate Research and Creative Works Symposium while the author was a graduate student at Rutgers University-Camden
Brain Control Interfaces in the coming age of Trans-humans: Philosophical and Regulatory Mappings
No full textThis article discusses the Philosophical and Regulatory mappings of Brain Control Interfaces in the coming age of Trans-humans. The author combines perspectives from Material Engagement theory, Don Ihde’s post-phenomenological perspective and Veerback’s cyborg intentionality on human-technology relations and juxtaposes this post-phenomenological hybrid-intentionality on upcoming BCI such as Neuralink. Taking the hybrid-intentionality as point of departure, the author then develops Regulatory approaches and principles in context of the normative harms borne out of such transfiguration while also highlighting these key harms
American Sign Language Recognition using Computer Vision and Deep Learning
No full textSign language recognition system helps to understand the hand gestures made by speech and hearing-impaired community that involves movement of fingers and different palm orientations. This framework has experienced significant growth in the field of computer vision and deep learning. The researcher has investigated various hardware and software approaches for accurate sign recognition. American sign language dataset was used to achieve the goal of this study. The author has performed exploratory data analysis to get insights into data and applied preprocessing techniques such as image resizing, smoothing, and re-scaling. Two deep learning models were implemented for this research, namely Convolutional Neural Network (CNN) and Residual Network 50 (ResNet50). A 2D CNN which consists of optimized hyper-parameters outperformed the other model and achieved an accuracy of 98.76%. A learning curve was also demonstrated for accuracy and loss which was considered during model evaluation
Modeling erythrocyte-released ATP in mouse aorta, and development of CAD models of vascular networks from experimental images
No full textThe methodology used to develop 3D CFD simulations to understand the role of erythrocyte or red blood cell (RBC)-released ATP on atherosclerotic plaque deposition is presented. A 3D non-planar CAD geometry featuring the entire length of a C57BL/6 mouse aorta is developed based on in vivo images and casts data. Simulation parameters based on in vivo data are set in COMSOL, and flow simulation based on Navier-Stokes equation is defined. ATP transport is modeled using an unsteady advection-diffusion-generation-degradation model where ATP is released from RBCs based on time-averaged shear rate. It was found that spatial distribution of WSS found through our simulation agrees with that of previous experimental and computational studies. Localized regions of ATP deposition found through our simulation match very well with the regions of plaque formation as observed in our in vivo study and other previous studies. Our model prediction and in vivo observations taken together suggest that shear-induced ATP release from RBCs deposits in localized regions of disturbed flow and contributes to the initiation of plaque formation in such regions. Various parametric studies are performed by varying baseline parameters to study their effects on ATP concentrations and spatial depositions. A secondary focus of this thesis is to present a methodology used to develop CAD models for high-fidelity modeling of microvascular blood flow. As such, generic techniques applicable to capillary vessel networks of any organ and vascular disfigurements are presented through examples of retinal microvascular and tumor microvascular networks. Meshing methodology using two different software is presented and techniques used to analyze them are described. A 1D mesh segment grouping method is presented which can be used to optimize the traditional postprocessing techniques.M.S.Includes bibliographical reference
Promotion of Entrepreneurship through Women-run Self-Help Groups
No full textSelf-help groups (SHGs), predominantly led by women, are anticipated to play a pivotal role in advancing gender equality, entrepreneurship, financial inclusion, and poverty alleviation within rural areas. Recognising this potential, both the Indian government and various state governments have implemented an array of policy interventions and initiatives over the years to promote livelihood generation and micro-entrepreneurship through SHGs. These initiatives encompass programs such as the SHG-Bank Linkage Programme by NABARD, DAY-NRLM, and various skill training programs. Despite the prevalence of SHGs actively involved in providing microfinance to their members, a significant number of SHG members do not engage in commercial activities, either collectively or individually. This reluctance persists despite the ongoing efforts made through policy interventions. This paper identifies the factors that motivate SHG members to embark on entrepreneurial endeavours and establishes correlations between these factors and the growth of SHGs. Researchers employed a mixed research methodology, conducting a survey among SHG members in the Narsinghpur district of Madhya Pradesh and conducting a policy analysis to identify existing gaps.
For the survey, a stratified random sampling technique was employed to identify specific SHG members within predefined strata. The research team conducted interviews of SHG members engaged in entrepreneurial activities and SHGs not engaged in any entrepreneurial activity. Statistical tools such as correlation coefficients were applied to analyse the survey data. Subsequently, the research findings were scrutinised to formulate recommendations for policy interventions aimed at more effectively promoting entrepreneurship among rural women
