@ksschool.org.in
Assistant Professor , K.S.School of Business management and Information Technology
Gujarat University
Computer Science, Artificial Intelligence, Information Systems, Computer Science Applications
Scopus Publications
Scholar Citations
Scholar h-index
Scholar i10-index
Nasrin Aasofwala, Shanti Verma, and Kalyani Patel
Springer Nature Singapore
Nasrin Aasofwala, Shanti Verma, and Kalyani Patel
IEEE
NLP (Natural Language Processing) is a use for process Hu-man Language. It is a combination of language and Artificial Intelligence. This paper is helpful to understand about social neuroscience, Translation, speech to text, NLP levels, challenges and Applications of NLP. Deaf and Dumb people of Gujarat use a Gujarati Language as a speech or sign language, therefore there is a need of translating English to Gujarati to help the people to understand the context of the language.Translation of study findings must be done carefully in order to enable the impact of neuroscientific insights on our everyday lives Deaf and Dumb people use Sign Language for communication, as Normal peopleuses a Speech So This paper includes various sign language and speech recognition module for converting Speech -> text -> sign. It includes Speech recognition API like apiai, assemblyai, google-cloud-speech, pocketsphinx, SpeechRecognition, watson-developer-cloud and wit. It represents the development and testing of speech to text and text to speech mechanism with python code. Translation of study findings must be done carefully in order to enable the impact of neuroscientific insights on our everyday lives. It implements English speech by speech recognition API and translates in a Gujarati text. This paper analyze the Precision, Accuracy and Recall on speech to text Datasets. It gives 96% Precision, 83% Recall and 83% Accuracy.
Nasrin Aasofwala, Shanti Verma, and Kalyani Patel
IEEE
Deaf Culture is important for deaf community as it is everywhere in the world. Deaf people are using Visual language (Sign language) for communicating. There are around 300 different types of sign languages are available in the globe like British Sign Language, Indonesian Sign Language, American sign language, etc. Each sign language has its own syntax and semantics. Some sign languages are using one hand gesture, some are using two hand gesture as they have their own rules for communication. There is a need of one standard form of sign language so it will be easier to understand. There are so many challenges and problems are facing by deaf community. Different sign languages are provided different solutions for speech to sign language and sign language to speech conversion. As there is no solution is provided by anyone for Gujarati Sign Language, we proposed a one communication model for Speech to Sign language. Speech will be recognized and convert into text, text will give the HamNoSys Notation (Sign language Notation) from a database and then it converts in SiGML format and then it display a sign animation (Avatar). That model will be helpful to Gujarat region deaf and dumb people for communicating with normal people.
Shanti Verma and Kalyani Patel
IEEE
The importance of the World Wide Web has increased enormously in recent years which leads to a large amount of information accessible through the web. The increasing importance of the web provides huge benefits to business, so to improve the business over web recommender systems has been proposed. Recommender systems works based on effective prediction which helps people for convenient access to their products that they might be interested in the real world. In this paper authors proposed a novel approach for product recommendation based on weighted product taxonomy. Product taxonomy is a hierarchical organization of products with different levels of hierarchy. Customer behavior and navigational factors are used for calculation of weight for product taxonomy. Authors also proposed a heuristic algorithm to search product “watch” in weighted product taxonomy. To prove that results of proposed heuristic algorithm are fast, authors apply independent sample ‘t’ test at 5% level of significance.
Shanti Verma and Kalyani Patel
IEEE
As we know number of smart phone users grows rapidly and Indian government survey said that there are 10+ billion smart phone users by 2020. We know that in metro cities people does not have time to purchase products of their daily needs. For this reason Online shopping using mobile phone is now in trend and survey also says that number of online shopping users grows exponentially. There are various types of factor plays important role in recommendation of product to customer. Product taxonomy, Customer behavioral and navigational factor are among them which are analyzed in this paper. Importance of weighted product taxonomy in fast recommendation of product is highlighted in this paper by authors. Authors proposed a greedy heuristic algorithm to search product in weighted product taxonomy. To proof the efficiency of proposed algorithm they use one tail independent sample‘t’ test with 5% level of significance and found that results are significant.
Meetu Kandpal and Kalyani Patel
IEEE
Success of any product may depend on the price of product. Demand of a product is one of the factors to be considered for deriving price of the product. As many IT companies have started to move towards the cloud computing and cloud resources are delivered as product over internet. There are many companies providing cloud services like salesforce.com, Amazon AWS, Microsoft azure etc. Different cloud service providers have different pricing policies to enhance the revenue and user satisfaction. The cloud providers have pricing schemes for cloud resources under fixed pricing and dynamic pricing. Some of them favor cloud providers, other cloud consumers. The paper presents a model to predict the price of cloud resource using Recurrent Neural Network(RNN) and auctioning method based on the parameters (as demand). The paper would give insight to researchers and cloud service providers to derive the policies based on the demand and other features.
Shanti Verma and Kalyani Patel
IEEE
In the era of Smartphone where users wants to perform any task in one click, the organization/planning of task is very important so that it can be done with minimum number of steps. Ontology is a method which is used in problem solving and which show product characteristics, properties and relationship between its entities. There are two ways to search; Text/keyword based and Ontology/semantic based. Ontology based search provides accurate and relevant information as compare to text based search because it will show only concise output of given query and also highlights the related keywords that are asked in query. Query uses the standard Meta data libraries like OWL and RDF. There are various algorithms available which can be used for Ontology based search. In this paper authors take the ontology of products available in Mobile Commerce as an example and try to find out the importance of Heuristic search for ontology and how it is helpful for predictive analysis and recommendation system. They also provide a comparative study of keyword and ontology based search and differences between Blind and Heuristic search.
Shanti Verma and Kalyani Patel
Springer Singapore
Shanti Verma and Kalyani Patel
IEEE
India has the second highest mobile phone users after China. In India using touch screen mobile device is fashion. All age group people now used touch screen handheld device to do shopping for daily need. Mobile commerce provides personalization and location based services to users. For those demographics of users plays an important role. In this paper authors try to find association between shopping habit and demographics of Indian mobile commerce users. For that authors conduct a online survey and collect data of 335 users. To conduct this experiment authors chose two demographics factors of Mobile commerce user's 1. Gender 2.yearly income. Through two way analysis of variance (ANOVA) we identified that 1) Gender and yearly income of customer are significant at 95% and 9999% confidence level respectively on shopping habit. 2) There is interaction between gender and yearly income towards shopping habit.
Shanti Verma and Kalyani Patel
IEEE
Mobile commerce users grows exponentially in India as India becomes number two in smart phone users in world. So there are lots of scopes to analyze mobile user's data to find unknown facts of Mobile commerce. To achieve objective of this paper authors conduct a survey named Mobile commerce usage in India. They collected a sample of 335 users in one month time duration via online medium (Google Forms). The major difference between online and traditional shopping is that in online shopping there is no touch, feel and trust. So the consumer gets afraid to pay first before receiving the product. In this paper authors try to find out relationship between type of product and payment method opted by Indian consumer in Mobile shopping. The result of experiment shows that P value is 0.005175 which is significant at 95% confidence. It tells that consumer opted payment method according to the type of product.
Meetu Kandpal, Monica Gahlawat, and Kalyani Patel
IEEE
In the era of Big Data analytics predictive modeling plays an important role to predict the future demand and behavior by using historical data. As majority of the IT companies running behind cloud services, the cloud service providers like Amazon, Google cloud, Microsoft Azure etc may be interested to know the future demand of the computing resources so that they can derive new pricing schemes to gain more profit. The providers have different pricing schemes to charge for computing resourcese. g., Amazon provides three pricing schemes, namely, on-demand pricing, reserved pricing and auction pricing in the same way Microsoft has different schemes like Pay-As-You-Go Subscriptions, Prepaid Subscriptions. The paper presents survey of role of predictive modeling in cloud service pricing. The survey result clearly shows that predictions made by various author are closer to actual outcomes, which highlights the importance of predictive modeling to forecast future demand of cloud computing resources and deciding the price of resources.
Kalyani A Patel and Jyoti S Pareek
IEEE
GH-MAP [6] is a rule based token mapping system for translation between sibling language pair Gujarati-Hindi. However, with this rule bases it is not possible to obtain 'most' appropriate translation in certain cases of post positions markers, pronouns, adjectives and adverbs, where post position markers, pronouns, adjectives and adverbs are under the influence of grammatical properties of other elements of the sentence. The translation problem of current GH-MAP has been resolved by enriching GH-MAP with special empirical rules, formed based on observed patterns. These empirical rules help GH-MAP to obtain more appropriate translation. The revamped GH-MAP has been evaluated on 332 Hindi sentences. It has proved the importance of empirical rules by achieving 4% improvement over an existing GH-MAP.