Pub Date : 2020-01-01DOI: 10.1109/Confluence47617.2020.9058016
Tushar Bhardwaj, Himanshu Upadhyay, S. Sharma
As the popularity of cloud computing is increasing there is an urgent requirement of developing highly efficient and highly qualitative cloud applications (CA). Hence, it be-comes a big research problem. A recommender system recommends the suitable item to the user and almost all the systems provide a rating score for preference. Traditionally, algorithms predicts the ratings that a user should give to the unrated components to queue the item in recommended list. To select an optimal candidate from a set of function-ally equivalent candidates is crucial through approaches that follow a framework for component quality ranking. More-over, such framework helps in detecting the poor performing candidates from a highly distributed cloud applications. In this paper a novel technique is proposed to provide personalized component ranking for designers by employing the past usage of components by different users. In this approach the similarity between the users is measured based on their rankings for functionally equivalent components set instead of their rating values. In this approach no additional invocation of cloud component is required. Experimental results on real world web-service invocations data set shows that the proposed approach outperforms the previous approaches.
{"title":"Framework for Quality Ranking of Components in Cloud Computing: Regressive Rank","authors":"Tushar Bhardwaj, Himanshu Upadhyay, S. Sharma","doi":"10.1109/Confluence47617.2020.9058016","DOIUrl":"https://doi.org/10.1109/Confluence47617.2020.9058016","url":null,"abstract":"As the popularity of cloud computing is increasing there is an urgent requirement of developing highly efficient and highly qualitative cloud applications (CA). Hence, it be-comes a big research problem. A recommender system recommends the suitable item to the user and almost all the systems provide a rating score for preference. Traditionally, algorithms predicts the ratings that a user should give to the unrated components to queue the item in recommended list. To select an optimal candidate from a set of function-ally equivalent candidates is crucial through approaches that follow a framework for component quality ranking. More-over, such framework helps in detecting the poor performing candidates from a highly distributed cloud applications. In this paper a novel technique is proposed to provide personalized component ranking for designers by employing the past usage of components by different users. In this approach the similarity between the users is measured based on their rankings for functionally equivalent components set instead of their rating values. In this approach no additional invocation of cloud component is required. Experimental results on real world web-service invocations data set shows that the proposed approach outperforms the previous approaches.","PeriodicalId":180005,"journal":{"name":"2020 10th International Conference on Cloud Computing, Data Science & Engineering (Confluence)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127551932","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-01-01DOI: 10.1109/Confluence47617.2020.9058137
Aishwarya, Parth Wadhwa, Prabhishek Singh
This research paper proposes an application of sentiment analysis that works on the principle of machine learning. The proposed application provides a comparative analysis of web series and movies of different genres of a particular time period on the basis of sentiments of the viewers. Data is fetched from twitter through API keys and twitter access tokens. The movies and web series from the year 2017 to 2019 of four different genres were taken and sentiment analysis was performed on each web series and movie, which gives result in the form of positive reviews and negative reviews. The famous hashtag for each movie and web series are determined. The total number of tweet counts is 3000. A Table of each genre was formed that contained the name of movie and web series, percentage of positive sentiments of corresponding web series or movie and percentage of negative sentiments of corresponding movie or web series. The graphical representation of each genre was done to analyze the results graphically. The combined analysis was performed after calculating the average percentage reviews of a positive and negative sentiment of all the movies and web series of each genre. The graphical representation of the combined analysis is done to analyze the final results. Through the proposed application results were analyzed concluding that whether movies or web series of a particular genre in the year 2017-19 were more liked by the viewers.
{"title":"A New Sentiment Analysis based Application for Analyzing Reviews of Web Series and Movies of Different Genres","authors":"Aishwarya, Parth Wadhwa, Prabhishek Singh","doi":"10.1109/Confluence47617.2020.9058137","DOIUrl":"https://doi.org/10.1109/Confluence47617.2020.9058137","url":null,"abstract":"This research paper proposes an application of sentiment analysis that works on the principle of machine learning. The proposed application provides a comparative analysis of web series and movies of different genres of a particular time period on the basis of sentiments of the viewers. Data is fetched from twitter through API keys and twitter access tokens. The movies and web series from the year 2017 to 2019 of four different genres were taken and sentiment analysis was performed on each web series and movie, which gives result in the form of positive reviews and negative reviews. The famous hashtag for each movie and web series are determined. The total number of tweet counts is 3000. A Table of each genre was formed that contained the name of movie and web series, percentage of positive sentiments of corresponding web series or movie and percentage of negative sentiments of corresponding movie or web series. The graphical representation of each genre was done to analyze the results graphically. The combined analysis was performed after calculating the average percentage reviews of a positive and negative sentiment of all the movies and web series of each genre. The graphical representation of the combined analysis is done to analyze the final results. Through the proposed application results were analyzed concluding that whether movies or web series of a particular genre in the year 2017-19 were more liked by the viewers.","PeriodicalId":180005,"journal":{"name":"2020 10th International Conference on Cloud Computing, Data Science & Engineering (Confluence)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124483598","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-01-01DOI: 10.1109/Confluence47617.2020.9057824
P. K. Upadhyay, Nadia Mohamed Kunhi, Y. Gupta, Shaik Ishraq Ahamed, Jidhin Das
Energy management is a vast subject of major significance and complexity. It entails in electing among the integrated set of sources to generate electrical energy that supplies to a set of loads by diminishing losses and expenses. The utilization of sources and consumption rate of loads are coherent, well-integrated and magnitude of the system, the optimal usage of sources must be performed in real-time to avoid power outage. With an increase in demands, there is an increase in improved productivity, which causes a reduction in greenhouse emissions and energy costs that are motivations for organizations to capitalize and implement new energy efficiency technologies and management strategies. This work aims to propose a system which can self-regulate a combined set of power sources namely green energy i.e., Solar-Wind Hybrid System and main grid, and loads organized as a unified group of individual systems, called micro-grid, to augment several measures such as cost-effectiveness and energy efficiency. This prototype is based on the multi-agent automated systems. These micro-grids, individually modelled as a self-directed entity, can interact and make its own decision giving the best outcome.
{"title":"Intelligent Energy Management System along with Solar-Wind Hybrid Power Source","authors":"P. K. Upadhyay, Nadia Mohamed Kunhi, Y. Gupta, Shaik Ishraq Ahamed, Jidhin Das","doi":"10.1109/Confluence47617.2020.9057824","DOIUrl":"https://doi.org/10.1109/Confluence47617.2020.9057824","url":null,"abstract":"Energy management is a vast subject of major significance and complexity. It entails in electing among the integrated set of sources to generate electrical energy that supplies to a set of loads by diminishing losses and expenses. The utilization of sources and consumption rate of loads are coherent, well-integrated and magnitude of the system, the optimal usage of sources must be performed in real-time to avoid power outage. With an increase in demands, there is an increase in improved productivity, which causes a reduction in greenhouse emissions and energy costs that are motivations for organizations to capitalize and implement new energy efficiency technologies and management strategies. This work aims to propose a system which can self-regulate a combined set of power sources namely green energy i.e., Solar-Wind Hybrid System and main grid, and loads organized as a unified group of individual systems, called micro-grid, to augment several measures such as cost-effectiveness and energy efficiency. This prototype is based on the multi-agent automated systems. These micro-grids, individually modelled as a self-directed entity, can interact and make its own decision giving the best outcome.","PeriodicalId":180005,"journal":{"name":"2020 10th International Conference on Cloud Computing, Data Science & Engineering (Confluence)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123766058","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-01-01DOI: 10.1109/Confluence47617.2020.9057985
Shrayan Mukhopadhyay, Shilpi Sharma
Biometrics are used to characterize a person's DNA, hand geometry, confront, and so on or behavioral qualities, for example, hand signature, voice tone, keystrokes et cetera. For that reason, these organic attributes are remarkable for each person. Much of the time, confront acknowledgment related advancements are winding up more mainstream among biometric-based advances that measure a person's regular information. Hereditary biometrics has, for the most part, used to validate and distinguish people by examining their physical attributes, for example, unique finger impression, eye iris, vein and so forth. Rather than utilizing a bank card, a camera presented at the ATM's would get pictures of countenances of clients, and separate them and the photographs of record holders in the database of banks to confirm the client's character. The motivation for driving this paper is to exhibit a Windows-based advancing application framework utilizing face certification checks.
{"title":"Real Time Facial Expression and Emotion Recognition using Eigen faces, LBPH and Fisher Algorithms","authors":"Shrayan Mukhopadhyay, Shilpi Sharma","doi":"10.1109/Confluence47617.2020.9057985","DOIUrl":"https://doi.org/10.1109/Confluence47617.2020.9057985","url":null,"abstract":"Biometrics are used to characterize a person's DNA, hand geometry, confront, and so on or behavioral qualities, for example, hand signature, voice tone, keystrokes et cetera. For that reason, these organic attributes are remarkable for each person. Much of the time, confront acknowledgment related advancements are winding up more mainstream among biometric-based advances that measure a person's regular information. Hereditary biometrics has, for the most part, used to validate and distinguish people by examining their physical attributes, for example, unique finger impression, eye iris, vein and so forth. Rather than utilizing a bank card, a camera presented at the ATM's would get pictures of countenances of clients, and separate them and the photographs of record holders in the database of banks to confirm the client's character. The motivation for driving this paper is to exhibit a Windows-based advancing application framework utilizing face certification checks.","PeriodicalId":180005,"journal":{"name":"2020 10th International Conference on Cloud Computing, Data Science & Engineering (Confluence)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122676651","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-01-01DOI: 10.1109/Confluence47617.2020.9057799
Mohammad Irfan Bala, M. Chishti
The information and communication technology has witnessed unprecedented changes with the introduction of IoT and the implementation of IoT applications is mainly dependant on cloud services like compute, storage, networking, etc. Fog computing has been introduced as a complement to the cloud infrastructure because in near future demands of the IoT devices will exceed the capabilities of the cloud. Our work focuses on the efficient utilization of the Cloud-Fog resources by distributing the application modules among Fog devices and cloud data centers. Placing the application modules on Fog devices improves performance parameters like response time, latency, energy consumption, etc. We have proposed two load balancing algorithms whose performance has been evaluated on the iFogSim simulator and their performance has been compared with the cloud-only approach. Our approach is generic which can be used in the vast majority of IoT applications.
{"title":"Offloading in Cloud and Fog Hybrid Infrastructure Using iFogSim","authors":"Mohammad Irfan Bala, M. Chishti","doi":"10.1109/Confluence47617.2020.9057799","DOIUrl":"https://doi.org/10.1109/Confluence47617.2020.9057799","url":null,"abstract":"The information and communication technology has witnessed unprecedented changes with the introduction of IoT and the implementation of IoT applications is mainly dependant on cloud services like compute, storage, networking, etc. Fog computing has been introduced as a complement to the cloud infrastructure because in near future demands of the IoT devices will exceed the capabilities of the cloud. Our work focuses on the efficient utilization of the Cloud-Fog resources by distributing the application modules among Fog devices and cloud data centers. Placing the application modules on Fog devices improves performance parameters like response time, latency, energy consumption, etc. We have proposed two load balancing algorithms whose performance has been evaluated on the iFogSim simulator and their performance has been compared with the cloud-only approach. Our approach is generic which can be used in the vast majority of IoT applications.","PeriodicalId":180005,"journal":{"name":"2020 10th International Conference on Cloud Computing, Data Science & Engineering (Confluence)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131535321","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-01-01DOI: 10.1109/Confluence47617.2020.9058109
Aayushi Mangal, Himanshu Malik, Garima Aggarwal
Data security being the main concern now a days, has faced a lot of threat in terms of breaching of information which requires immediate attention. Biometrics have served a long-run for this purpose which is a part of Deep Learning. In the recent past, face recognition has become a very important tool for safety and security purposes. This paper presents the application of face recognition technique, making use of Convolutional Neural Network (CNN) with Python and a comparison is drawn between the other techniques such as Principal Component Analysis (PCA), Local Binary Pattern (LBP) and K Nearest Neighbour (KNN). Unlike conventional methods, the proposed scheme uses four Convolutional layers with ReLu layers, four pooling layers, a fully connected layer and a Softmax Loss Layer to normalize the probability distribution. The dataset consists of 1500 images with different facial expressions and the model is trained and tested in order to acquire an accuracy using CNN method. Experimental results show that the proposed Neural Network scored an accuracy of 96.96%.
{"title":"An Efficient Convolutional Neural Network Approach for Facial Recognition","authors":"Aayushi Mangal, Himanshu Malik, Garima Aggarwal","doi":"10.1109/Confluence47617.2020.9058109","DOIUrl":"https://doi.org/10.1109/Confluence47617.2020.9058109","url":null,"abstract":"Data security being the main concern now a days, has faced a lot of threat in terms of breaching of information which requires immediate attention. Biometrics have served a long-run for this purpose which is a part of Deep Learning. In the recent past, face recognition has become a very important tool for safety and security purposes. This paper presents the application of face recognition technique, making use of Convolutional Neural Network (CNN) with Python and a comparison is drawn between the other techniques such as Principal Component Analysis (PCA), Local Binary Pattern (LBP) and K Nearest Neighbour (KNN). Unlike conventional methods, the proposed scheme uses four Convolutional layers with ReLu layers, four pooling layers, a fully connected layer and a Softmax Loss Layer to normalize the probability distribution. The dataset consists of 1500 images with different facial expressions and the model is trained and tested in order to acquire an accuracy using CNN method. Experimental results show that the proposed Neural Network scored an accuracy of 96.96%.","PeriodicalId":180005,"journal":{"name":"2020 10th International Conference on Cloud Computing, Data Science & Engineering (Confluence)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131728431","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-01-01DOI: 10.1109/Confluence47617.2020.9058152
R. Bhimanpallewar, M. R. Narasingarao
Machine learning is one of the recent trends. Currently it is being used in variety of interdisciplinary domains. The major contribution of GDP (Gross Domestic Product) of India belongs to agriculture production directly or indirectly. Most of the population in India is still dependent on farming or livestock for their regular income. Due to sufficient availability of solar energy, gifted by nature, in India we found the variety of crops. Major farmers hold fragmented land and adapt rain-feed cropping with traditional and repeated crop pattern. For increasing yield farmers add the fertilizers in extra quantity, which leads to soil degradation. Rather than repeated crop farmer should go for suitable crops, according available environmental condition. Here we have discussed machine learning approaches to develop Agriculture Advisory System. Comparative analysis of different supervised techniques with hybrid approach is done with the help of their performances.
{"title":"Alternative approaches of Machine Learning for Agriculture Advisory System","authors":"R. Bhimanpallewar, M. R. Narasingarao","doi":"10.1109/Confluence47617.2020.9058152","DOIUrl":"https://doi.org/10.1109/Confluence47617.2020.9058152","url":null,"abstract":"Machine learning is one of the recent trends. Currently it is being used in variety of interdisciplinary domains. The major contribution of GDP (Gross Domestic Product) of India belongs to agriculture production directly or indirectly. Most of the population in India is still dependent on farming or livestock for their regular income. Due to sufficient availability of solar energy, gifted by nature, in India we found the variety of crops. Major farmers hold fragmented land and adapt rain-feed cropping with traditional and repeated crop pattern. For increasing yield farmers add the fertilizers in extra quantity, which leads to soil degradation. Rather than repeated crop farmer should go for suitable crops, according available environmental condition. Here we have discussed machine learning approaches to develop Agriculture Advisory System. Comparative analysis of different supervised techniques with hybrid approach is done with the help of their performances.","PeriodicalId":180005,"journal":{"name":"2020 10th International Conference on Cloud Computing, Data Science & Engineering (Confluence)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130432363","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-01-01DOI: 10.1109/Confluence47617.2020.9057949
Surbhi Vijh, Deepak Gaur, Sushil Kumar
Hypertension is named as silent killer. It is considered as one of alarming factor for chronic kidney disease, heart failure, impaired vision, Ischemic heart disease, Stroke etc. Hypertension is divided into systolic and diastolic blood pressure. According to studies 90-95% cause of hypertension is change in lifestyle therefore Diet plays essential role to hypertension patient. According to WHO studies, Death due to chronic disease in increased by 18% in India. However high blood pressure had affected 1.13 billion people across the world. The observed systolic blood pressure measurement is > 140 mmHg and diastolic blood pressure measurement is > 90mmHg in 2015. The paper shows the finest diet plan for hypertension patient using Analytic Hierarchy process. The technique used in this paper for representing diet plan is unique and haven’t been shown earlier. The Diet plan considers all the meals needed to be consumed by hypertension patient in breakfast, lunch and dinner. The results are validated using Entropy method. The results evaluated during validation are same as obtained using AHP.
{"title":"Diet Recommendation for Hypertension Patient on basis of Nutrient using AHP and Entropy","authors":"Surbhi Vijh, Deepak Gaur, Sushil Kumar","doi":"10.1109/Confluence47617.2020.9057949","DOIUrl":"https://doi.org/10.1109/Confluence47617.2020.9057949","url":null,"abstract":"Hypertension is named as silent killer. It is considered as one of alarming factor for chronic kidney disease, heart failure, impaired vision, Ischemic heart disease, Stroke etc. Hypertension is divided into systolic and diastolic blood pressure. According to studies 90-95% cause of hypertension is change in lifestyle therefore Diet plays essential role to hypertension patient. According to WHO studies, Death due to chronic disease in increased by 18% in India. However high blood pressure had affected 1.13 billion people across the world. The observed systolic blood pressure measurement is > 140 mmHg and diastolic blood pressure measurement is > 90mmHg in 2015. The paper shows the finest diet plan for hypertension patient using Analytic Hierarchy process. The technique used in this paper for representing diet plan is unique and haven’t been shown earlier. The Diet plan considers all the meals needed to be consumed by hypertension patient in breakfast, lunch and dinner. The results are validated using Entropy method. The results evaluated during validation are same as obtained using AHP.","PeriodicalId":180005,"journal":{"name":"2020 10th International Conference on Cloud Computing, Data Science & Engineering (Confluence)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123381827","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-01-01DOI: 10.1109/Confluence47617.2020.9058286
Vishesh Kasturia, Shanu Sharma, Sachin Sharma
From past few decades information technology industry is on the rise and software development companies thrive to provide the best results for the consumers. Sentiment Analysis is a powerful tool that can help the software industry and company to better evaluate user needs and cater the software in a way to maximise the sales potential. Sentiment Analysis combined with machine learning techniques can help us learn about the industrial trends. Greater than 40 thousand Exabyte (10^18) of data is estimated to be a part of the internet out of which 80% is unstructured and can be processed to useful means using NLP techniques. In proposed work sentiment analysis has been applied on user review to predict its saleability or in simpler words: How well a product will sell? Customer feedback was collected from users through a feedback form which required them to express their satisfaction with the product by answering a set of questions which serves as features and input to the machine which evaluates the features such as user interface, Performance, feasibility, cost effectiveness and customer service by extracting the polarity from each. The result shows that sentiment analysis is a viable option to predict the saleability of a product. The empirical results are close to the customer’s own expected probability of buying.
{"title":"Automatic Product Saleability Prediction using Sentiment Analysis on User Reviews","authors":"Vishesh Kasturia, Shanu Sharma, Sachin Sharma","doi":"10.1109/Confluence47617.2020.9058286","DOIUrl":"https://doi.org/10.1109/Confluence47617.2020.9058286","url":null,"abstract":"From past few decades information technology industry is on the rise and software development companies thrive to provide the best results for the consumers. Sentiment Analysis is a powerful tool that can help the software industry and company to better evaluate user needs and cater the software in a way to maximise the sales potential. Sentiment Analysis combined with machine learning techniques can help us learn about the industrial trends. Greater than 40 thousand Exabyte (10^18) of data is estimated to be a part of the internet out of which 80% is unstructured and can be processed to useful means using NLP techniques. In proposed work sentiment analysis has been applied on user review to predict its saleability or in simpler words: How well a product will sell? Customer feedback was collected from users through a feedback form which required them to express their satisfaction with the product by answering a set of questions which serves as features and input to the machine which evaluates the features such as user interface, Performance, feasibility, cost effectiveness and customer service by extracting the polarity from each. The result shows that sentiment analysis is a viable option to predict the saleability of a product. The empirical results are close to the customer’s own expected probability of buying.","PeriodicalId":180005,"journal":{"name":"2020 10th International Conference on Cloud Computing, Data Science & Engineering (Confluence)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115067148","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-01-01DOI: 10.1109/Confluence47617.2020.9058207
S. Bali, S. S. Tyagi
Deep learning is the key technology used in a large variety of applications such as self-driving cars, image recognition, automatic machine translation, automatic handwriting generation. The success was fueled due to accessibility of huge datasets, GPUs, max pooling. Earlier machine learning techniques employed two phases: features extraction and classification. The performance of such algorithms was highly dependent on how well the features are extracted and that was the major bottleneck of these techniques. Deep learning techniques employ Convolutional Neural Networks (CNNs) with numerous layers of non-linear processing for extracting the features automatically and classification that solves the previous problem. In the real time applications most of the time, either the dataset is unavailable or has less amount of data which makes it difficult to achieve accurate results for classifying the images. CNNs are hard to be trained using the small datasets. Transfer learning has emerged as a very powerful technique where in the knowledge gained from the larger dataset is transferred to the new dataset. Data augmentation and dropout are also powerful techniques that are useful for dealing with small datasets. In this paper, different techniques using the VGG16 pretrained model are compared on the small dataset.
{"title":"Evaluation of transfer learning techniques for classifying small surgical dataset","authors":"S. Bali, S. S. Tyagi","doi":"10.1109/Confluence47617.2020.9058207","DOIUrl":"https://doi.org/10.1109/Confluence47617.2020.9058207","url":null,"abstract":"Deep learning is the key technology used in a large variety of applications such as self-driving cars, image recognition, automatic machine translation, automatic handwriting generation. The success was fueled due to accessibility of huge datasets, GPUs, max pooling. Earlier machine learning techniques employed two phases: features extraction and classification. The performance of such algorithms was highly dependent on how well the features are extracted and that was the major bottleneck of these techniques. Deep learning techniques employ Convolutional Neural Networks (CNNs) with numerous layers of non-linear processing for extracting the features automatically and classification that solves the previous problem. In the real time applications most of the time, either the dataset is unavailable or has less amount of data which makes it difficult to achieve accurate results for classifying the images. CNNs are hard to be trained using the small datasets. Transfer learning has emerged as a very powerful technique where in the knowledge gained from the larger dataset is transferred to the new dataset. Data augmentation and dropout are also powerful techniques that are useful for dealing with small datasets. In this paper, different techniques using the VGG16 pretrained model are compared on the small dataset.","PeriodicalId":180005,"journal":{"name":"2020 10th International Conference on Cloud Computing, Data Science & Engineering (Confluence)","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2020-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129562386","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：