This paper develops a procedure to "recover" the missing data of a personal accounting application. The missing data are estimated using a thesaurus matching method and a neural network model. The data sets are split into two parts, the expenditure data and the income data. To estimate the users' missing expenditure data, this paper uses a thesaurus matching method combined with text segmentation technology, successfully reclassifying the accounting data and mining the users' accounting habits. In order to infer the almost vacant income data inversely from the users' expenditure data, a neural network is trained to deduct the relationship between expenditure data and income data, using the income and expenditure sample data of 20,133 households mined from Chinese Household Financial Survey (CHFS) database. The recovered accounting data would be helpful for IT companies in analyzing users' consumption habits and income status, building users' portraits and designing personalized investment products for users. Finally, after dividing users into four categories based on clustering algorithm, the types and quantity of investment products are designed for each group of users to optimize users' asset allocation structures and to make advertisements targeted.
{"title":"User Portraits and Investment Planning Based on Accounting Data","authors":"Yibing Wu, Rongxuan Wang, Wei Dai, Shixuan Dong, Xiaohe You, Huanxiong You, Lijie Liu","doi":"10.1109/SCC49832.2020.00059","DOIUrl":"https://doi.org/10.1109/SCC49832.2020.00059","url":null,"abstract":"This paper develops a procedure to \"recover\" the missing data of a personal accounting application. The missing data are estimated using a thesaurus matching method and a neural network model. The data sets are split into two parts, the expenditure data and the income data. To estimate the users' missing expenditure data, this paper uses a thesaurus matching method combined with text segmentation technology, successfully reclassifying the accounting data and mining the users' accounting habits. In order to infer the almost vacant income data inversely from the users' expenditure data, a neural network is trained to deduct the relationship between expenditure data and income data, using the income and expenditure sample data of 20,133 households mined from Chinese Household Financial Survey (CHFS) database. The recovered accounting data would be helpful for IT companies in analyzing users' consumption habits and income status, building users' portraits and designing personalized investment products for users. Finally, after dividing users into four categories based on clustering algorithm, the types and quantity of investment products are designed for each group of users to optimize users' asset allocation structures and to make advertisements targeted.","PeriodicalId":274909,"journal":{"name":"2020 IEEE International Conference on Services Computing (SCC)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130483651","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 : 2020-11-01DOI: 10.1109/SCC49832.2020.00030
Xue Han, L. Hu, J. Sen, Yabin Dang, Buyu Gao, Vatche Isahagian, Chuan Lei, Vasilis Efthymiou, Fatma Özcan, A. Quamar, Ziming Huang, Vinod Muthusamy
Advances in the adoption of business process management platforms have led to increasing volumes runtime event logs, containing information about the execution of the process. Business users analyze this event data for real-time insights on performance and optimization opportunities. However, querying the event data is difficult for business users without knowing the details of the backend store, data schema, and query languages. Consequently, the business insights are mostly limited to static dashboards, only capturing predefined performance metrics. In this paper, we introduce an interface for business users to query the business event data using natural language, without knowing the exact schema of the event data or the query language. Moreover, we propose a bootstrapping pipeline, which utilizes both event data and business domain-specific artifacts to automatically instantiate the natural language interface over the event data. We build and evaluate our prototype over datasets from both practical projects and public challenge events data stored in Elasticsearch. Experimental results show that our system produces an average accuracy of 80% across all data sets, with high precision ( 91%) and good recall ( 81%).
{"title":"Bootstrapping Natural Language Querying on Process Automation Data","authors":"Xue Han, L. Hu, J. Sen, Yabin Dang, Buyu Gao, Vatche Isahagian, Chuan Lei, Vasilis Efthymiou, Fatma Özcan, A. Quamar, Ziming Huang, Vinod Muthusamy","doi":"10.1109/SCC49832.2020.00030","DOIUrl":"https://doi.org/10.1109/SCC49832.2020.00030","url":null,"abstract":"Advances in the adoption of business process management platforms have led to increasing volumes runtime event logs, containing information about the execution of the process. Business users analyze this event data for real-time insights on performance and optimization opportunities. However, querying the event data is difficult for business users without knowing the details of the backend store, data schema, and query languages. Consequently, the business insights are mostly limited to static dashboards, only capturing predefined performance metrics. In this paper, we introduce an interface for business users to query the business event data using natural language, without knowing the exact schema of the event data or the query language. Moreover, we propose a bootstrapping pipeline, which utilizes both event data and business domain-specific artifacts to automatically instantiate the natural language interface over the event data. We build and evaluate our prototype over datasets from both practical projects and public challenge events data stored in Elasticsearch. Experimental results show that our system produces an average accuracy of 80% across all data sets, with high precision ( 91%) and good recall ( 81%).","PeriodicalId":274909,"journal":{"name":"2020 IEEE International Conference on Services Computing (SCC)","volume":"44 10","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"120903916","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 : 2020-11-01DOI: 10.1109/SCC49832.2020.00032
Zhongjie Wang, Xiang He, Lei Liu, Zhiying Tu, Hanchuan Xu
In software engineering research, software evolution is always a hot focus. A dominating driving force of software evolution is requirement changes (RCs). In this paper, we make a comprehensive survey on start-of-the-art progress of requirement-driven software evolution, especially aiming at microservice-based software systems (MSS). MSS has become a dominating architecture style for modern software because of its advantage on agile DevOps and superior supports on business agility, thus it has been proved to outperform other architecture styles on fitting for requirement changes. A high-level conceptual framework for requirement-driven MSS evolution is demonstrated first, then related work are surveyed in terms of sources, representations and types of RCs, approaches for capturing RCs and mapping them to MSS evolution, and various techniques for MSS evolution in microservice, architecture, and infrastructure levels, respectively. Limitations of existing works are discussed and potential research topics are listed for reference. An integrated platform supporting full-lifecycle requirement-driven MSS evolution is introduced at last. We do hope this survey would help researchers strive for deep insights in this topic.
{"title":"Survey on Requirement-Driven Microservice System Evolution","authors":"Zhongjie Wang, Xiang He, Lei Liu, Zhiying Tu, Hanchuan Xu","doi":"10.1109/SCC49832.2020.00032","DOIUrl":"https://doi.org/10.1109/SCC49832.2020.00032","url":null,"abstract":"In software engineering research, software evolution is always a hot focus. A dominating driving force of software evolution is requirement changes (RCs). In this paper, we make a comprehensive survey on start-of-the-art progress of requirement-driven software evolution, especially aiming at microservice-based software systems (MSS). MSS has become a dominating architecture style for modern software because of its advantage on agile DevOps and superior supports on business agility, thus it has been proved to outperform other architecture styles on fitting for requirement changes. A high-level conceptual framework for requirement-driven MSS evolution is demonstrated first, then related work are surveyed in terms of sources, representations and types of RCs, approaches for capturing RCs and mapping them to MSS evolution, and various techniques for MSS evolution in microservice, architecture, and infrastructure levels, respectively. Limitations of existing works are discussed and potential research topics are listed for reference. An integrated platform supporting full-lifecycle requirement-driven MSS evolution is introduced at last. We do hope this survey would help researchers strive for deep insights in this topic.","PeriodicalId":274909,"journal":{"name":"2020 IEEE International Conference on Services Computing (SCC)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132148060","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 : 2020-11-01DOI: 10.1109/SCC49832.2020.00028
Marwa Elleuch, Nour Assy, N. Laga, Walid Gaaloul, Oumaima Alaoui Ismaili, B. Benatallah
Significant research work has been conducted in the area of process mining leading to mature solutions for discovering knowledge from structured process event logs analysis. Recently, there were several initiatives to extend the scope of these analysis to consider heterogeneous and unstructured data sources. More precisely, email analysis has attracted much attention as emailing system is considered as the principal channel to support the execution of business processes. However, existing initiatives didn’t formalize the relationship between emailing systems and business process elements. As a result, they target to discover business processes limited to the activity perspective. In this paper, we first propose a meta model to specify what kind of process knowledge we can discover from emails. We define by this way a research roadmap for an effective multi-perspective process discovery from emails. This metamodel is proved through a concrete case study related to "hiring", "patent application", and "paper submission" business processes. In addition, we highlight the limitations of current process mining techniques in the discovery of different process perspectives.
{"title":"A Meta Model for Mining Processes from Email Data","authors":"Marwa Elleuch, Nour Assy, N. Laga, Walid Gaaloul, Oumaima Alaoui Ismaili, B. Benatallah","doi":"10.1109/SCC49832.2020.00028","DOIUrl":"https://doi.org/10.1109/SCC49832.2020.00028","url":null,"abstract":"Significant research work has been conducted in the area of process mining leading to mature solutions for discovering knowledge from structured process event logs analysis. Recently, there were several initiatives to extend the scope of these analysis to consider heterogeneous and unstructured data sources. More precisely, email analysis has attracted much attention as emailing system is considered as the principal channel to support the execution of business processes. However, existing initiatives didn’t formalize the relationship between emailing systems and business process elements. As a result, they target to discover business processes limited to the activity perspective. In this paper, we first propose a meta model to specify what kind of process knowledge we can discover from emails. We define by this way a research roadmap for an effective multi-perspective process discovery from emails. This metamodel is proved through a concrete case study related to \"hiring\", \"patent application\", and \"paper submission\" business processes. In addition, we highlight the limitations of current process mining techniques in the discovery of different process perspectives.","PeriodicalId":274909,"journal":{"name":"2020 IEEE International Conference on Services Computing (SCC)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133012006","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 : 2020-11-01DOI: 10.1109/SCC49832.2020.00046
Zhaoning Wang, B. Cheng, Wenkai Zhang, Junliang Chen
With the evolution of web technologies, various services become available in the pervasive network environment. Combining atomic services via the input and output dependency according to functional requirements with the multiple nonfunctional Quality-of-Service (QoS) guarantees has become a widely considered optimization problem. The conventional multi-objective service composition relying on manually predefined service chains fails to ensure global optimality. Although the automatic service composition successfully expands the search space, the searching graph which it relies on causes computationally expensive and fails to handle multiple objectives. Therefore, this paper proposes a novel efficient multi-objective automatic service composition approach. Particularly, it introduces a service execution timeline model to decompose the composition problem into several sub-problems to reduce computational complexity. Further, it employs an evolutionary process to explore the search space and determine the approximately Pareto front of the composition solutions. The experimental results on the benchmarks show that our approach could achieve a better trade-off between the computation cost and ensuring a better QoS compared with two recently proposed automatic composition approaches.
{"title":"QoS-aware Automatic Service Composition Based on Service Execution Timeline with Multi-objective Optimization","authors":"Zhaoning Wang, B. Cheng, Wenkai Zhang, Junliang Chen","doi":"10.1109/SCC49832.2020.00046","DOIUrl":"https://doi.org/10.1109/SCC49832.2020.00046","url":null,"abstract":"With the evolution of web technologies, various services become available in the pervasive network environment. Combining atomic services via the input and output dependency according to functional requirements with the multiple nonfunctional Quality-of-Service (QoS) guarantees has become a widely considered optimization problem. The conventional multi-objective service composition relying on manually predefined service chains fails to ensure global optimality. Although the automatic service composition successfully expands the search space, the searching graph which it relies on causes computationally expensive and fails to handle multiple objectives. Therefore, this paper proposes a novel efficient multi-objective automatic service composition approach. Particularly, it introduces a service execution timeline model to decompose the composition problem into several sub-problems to reduce computational complexity. Further, it employs an evolutionary process to explore the search space and determine the approximately Pareto front of the composition solutions. The experimental results on the benchmarks show that our approach could achieve a better trade-off between the computation cost and ensuring a better QoS compared with two recently proposed automatic composition approaches.","PeriodicalId":274909,"journal":{"name":"2020 IEEE International Conference on Services Computing (SCC)","volume":"36 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124898605","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 : 2020-11-01DOI: 10.1109/SCC49832.2020.00068
Yongshen Long, Wuqiao Chen, Xutao Li, Yunming Ye
In China, many government platforms begin to offer interfaces to each other for establishing urban service collaboration systems. However, it is expensive and tedious to migrate a successful collaboration procedure from one city to another as there are no uniform standards on API definitions. In this paper, we aim to develop a method that can match the cross-city APIs for service collaboration migration. We consider the matching task as a binary classification problem. A semantic feature engineering scheme is proposed and the matching is achieved via an XGBoost classifier. Experiments demonstrate the effectiveness of the proposed method.
{"title":"Automatic Cross-City API Matching for Urban Service Collaboration Based on Semantics","authors":"Yongshen Long, Wuqiao Chen, Xutao Li, Yunming Ye","doi":"10.1109/SCC49832.2020.00068","DOIUrl":"https://doi.org/10.1109/SCC49832.2020.00068","url":null,"abstract":"In China, many government platforms begin to offer interfaces to each other for establishing urban service collaboration systems. However, it is expensive and tedious to migrate a successful collaboration procedure from one city to another as there are no uniform standards on API definitions. In this paper, we aim to develop a method that can match the cross-city APIs for service collaboration migration. We consider the matching task as a binary classification problem. A semantic feature engineering scheme is proposed and the matching is achieved via an XGBoost classifier. Experiments demonstrate the effectiveness of the proposed method.","PeriodicalId":274909,"journal":{"name":"2020 IEEE International Conference on Services Computing (SCC)","volume":"102 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127195035","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 : 2020-11-01DOI: 10.1109/SCC49832.2020.00053
Yincheng Lou, Yanmei Zhang, Shiping Chen
As one of the leading blockchain systems in operation, Ethereum has numerous smart contracts deployed to implement a variety of functions. Unfortunately, speculators introduce scams such as Ponzi scheme in the traditional financial sector into some of these smart contracts, causing millions of dollars of losses to investors. At present, there are a few of quantitative identification methods for new fraud modes under the background of Internet finance, and detection methods for the Ponzi scheme contracts on Ethereum are even less. In this paper, we propose an improved convolutional neural network as a detection model for Ponzi schemes in smart contracts. We use real smart contracts to evaluate the feasibility and usefulness of our mode. Results show that our improved convolutional neural network can overcome difficulties in training caused by different length of smart contracts' bytecodes. Compared with the state-of-the-art methods, the precision and recall rate of our model for Ponzi scheme detection are improved by 3.2% and 24.8% respectively.
{"title":"Ponzi Contracts Detection Based on Improved Convolutional Neural Network","authors":"Yincheng Lou, Yanmei Zhang, Shiping Chen","doi":"10.1109/SCC49832.2020.00053","DOIUrl":"https://doi.org/10.1109/SCC49832.2020.00053","url":null,"abstract":"As one of the leading blockchain systems in operation, Ethereum has numerous smart contracts deployed to implement a variety of functions. Unfortunately, speculators introduce scams such as Ponzi scheme in the traditional financial sector into some of these smart contracts, causing millions of dollars of losses to investors. At present, there are a few of quantitative identification methods for new fraud modes under the background of Internet finance, and detection methods for the Ponzi scheme contracts on Ethereum are even less. In this paper, we propose an improved convolutional neural network as a detection model for Ponzi schemes in smart contracts. We use real smart contracts to evaluate the feasibility and usefulness of our mode. Results show that our improved convolutional neural network can overcome difficulties in training caused by different length of smart contracts' bytecodes. Compared with the state-of-the-art methods, the precision and recall rate of our model for Ponzi scheme detection are improved by 3.2% and 24.8% respectively.","PeriodicalId":274909,"journal":{"name":"2020 IEEE International Conference on Services Computing (SCC)","volume":"17 3","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"120993185","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}
In recent years, various innovative service patterns have been practiced with the blooming of modern service industry (MSI). However, its theoretical construction lags far behind practice. The existing service models in concept-level or process languages in practice-level cannot fully express the connotation of service pattern, which covers concept, modeling, and practice, bringing about the difficulties on pattern modeling and simulation. This paper proposes a service pattern description model which decouples the service pattern into two layers: the element layer and the relation layer. Meanwhile, a process-oriented closed loop simulation framework is developed and applied on the Rural Taobao, a case of Alibaba. The simulation framework contains four parts, namely market environment, process calculate, value output and regulatory environment. In case study, we illustrate that our methods can help business analyst identify bottlenecks and estimate the performance of service pattern. At last, the ability of our methods are discussed by comparing with other models and simulation approaches.
{"title":"Service Pattern Modeling and Simulation: A Case Study of Rural Taobao","authors":"Jintao Chen, Jianwei Yin, Meng Xi, Siwei Tan, Yongna Wei, Shuiguang Deng","doi":"10.1109/SCC49832.2020.00012","DOIUrl":"https://doi.org/10.1109/SCC49832.2020.00012","url":null,"abstract":"In recent years, various innovative service patterns have been practiced with the blooming of modern service industry (MSI). However, its theoretical construction lags far behind practice. The existing service models in concept-level or process languages in practice-level cannot fully express the connotation of service pattern, which covers concept, modeling, and practice, bringing about the difficulties on pattern modeling and simulation. This paper proposes a service pattern description model which decouples the service pattern into two layers: the element layer and the relation layer. Meanwhile, a process-oriented closed loop simulation framework is developed and applied on the Rural Taobao, a case of Alibaba. The simulation framework contains four parts, namely market environment, process calculate, value output and regulatory environment. In case study, we illustrate that our methods can help business analyst identify bottlenecks and estimate the performance of service pattern. At last, the ability of our methods are discussed by comparing with other models and simulation approaches.","PeriodicalId":274909,"journal":{"name":"2020 IEEE International Conference on Services Computing (SCC)","volume":"150 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123064631","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 : 2020-11-01DOI: 10.1109/SCC49832.2020.00021
Benjamin A. Kwapong, R. Anarfi, K. K. Fletcher
Tags have been extensively used to organize and index mashup services. However, the selection of relevant tags that depict functionality of mashups has remained a daunting task. This is because mashups have different functionalities than their constituent web APIs. Some existing tag recommendation methods usually follow a manual approach, which is time consuming and prone to errors. Others propose some means of automatic tag recommendation that use a similarity measure which has to be re-computed for every new mashup against the entire mashup and web API database. Such methods are also time consuming, inefficient and therefore not practical. In this paper, we present an automatic tag recommendation method for mashups, using knowledge graphs (KG). The method uses as entry points (seeds) into the KG, topics from mashup description, its primary category, and its constituent web APIs. From the seeds, we walk the graph to extract candidate tags based on node cosine similarity. We finally employ word similarity as a scoring function to explore and rank the candidate tags. Top-ranked candidate tags are subsequently recommended. We conduct experiments, with a real world dataset from programmable web1, and compare our results to existing baselines. Our results show that our model outperforms the baselines in all cases.
{"title":"A Knowledge Graph Approach to Mashup Tag Recommendation","authors":"Benjamin A. Kwapong, R. Anarfi, K. K. Fletcher","doi":"10.1109/SCC49832.2020.00021","DOIUrl":"https://doi.org/10.1109/SCC49832.2020.00021","url":null,"abstract":"Tags have been extensively used to organize and index mashup services. However, the selection of relevant tags that depict functionality of mashups has remained a daunting task. This is because mashups have different functionalities than their constituent web APIs. Some existing tag recommendation methods usually follow a manual approach, which is time consuming and prone to errors. Others propose some means of automatic tag recommendation that use a similarity measure which has to be re-computed for every new mashup against the entire mashup and web API database. Such methods are also time consuming, inefficient and therefore not practical. In this paper, we present an automatic tag recommendation method for mashups, using knowledge graphs (KG). The method uses as entry points (seeds) into the KG, topics from mashup description, its primary category, and its constituent web APIs. From the seeds, we walk the graph to extract candidate tags based on node cosine similarity. We finally employ word similarity as a scoring function to explore and rank the candidate tags. Top-ranked candidate tags are subsequently recommended. We conduct experiments, with a real world dataset from programmable web1, and compare our results to existing baselines. Our results show that our model outperforms the baselines in all cases.","PeriodicalId":274909,"journal":{"name":"2020 IEEE International Conference on Services Computing (SCC)","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132596868","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 : 2020-11-01DOI: 10.1109/SCC49832.2020.00020
Bayan I. Alghofaily, Chen Ding
In this work, we propose a unique approach for Machine Learning (ML) service recommendation using multilayer perceptron architecture. A service is recommended based on its predicted performance on the input dataset. We take Quality of Services (QoS) as the performance indicator. Depending on the application domain and user requirements, the importance level of different QoS attributes could be different. For ML services, their QoS values are affected by both the input dataset and the service. It would be helpful if we can include their features into the recommendation model. In this work, we consider two types of side information: features of the services and of the user (in our case the dataset given by the user). In the experiment, we take OpenML as our data source and extract QoS values of multiple classification services running on 390 datasets. The result shows that dataset-service interactions can be used to predict the performance of a service on a given dataset. When we integrate all the side information, the performance is better than using the interaction data alone in terms of both prediction and recommendation accuracy.
{"title":"MLP4ML: Machine Learning Service Recommendation System using MLP","authors":"Bayan I. Alghofaily, Chen Ding","doi":"10.1109/SCC49832.2020.00020","DOIUrl":"https://doi.org/10.1109/SCC49832.2020.00020","url":null,"abstract":"In this work, we propose a unique approach for Machine Learning (ML) service recommendation using multilayer perceptron architecture. A service is recommended based on its predicted performance on the input dataset. We take Quality of Services (QoS) as the performance indicator. Depending on the application domain and user requirements, the importance level of different QoS attributes could be different. For ML services, their QoS values are affected by both the input dataset and the service. It would be helpful if we can include their features into the recommendation model. In this work, we consider two types of side information: features of the services and of the user (in our case the dataset given by the user). In the experiment, we take OpenML as our data source and extract QoS values of multiple classification services running on 390 datasets. The result shows that dataset-service interactions can be used to predict the performance of a service on a given dataset. When we integrate all the side information, the performance is better than using the interaction data alone in terms of both prediction and recommendation accuracy.","PeriodicalId":274909,"journal":{"name":"2020 IEEE International Conference on Services Computing (SCC)","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131514387","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：