No abstract

Service-oriented software engineering is a software engineering methodology focussed on the development of software systems. The systematic application of technological and scientific knowledge depends on the methodology, experience, design for obtaining efficient implementation, testing and software documentation. Software effort estimation (SEE) plays an essential role in reusable service for ensembling the effort estimation of the software development. Effort estimation is the most efficient process applied in software engineering for the prediction of effort. SEE methods are utilised to achieve the effort, cost and human resources with the assistance of the dataset. It is hard to predict the cost, effort, size and schedule consistently through SEE and hence it causes damage to software enterprises. To overwhelm these limitations, an enhanced support vector regression algorithm is used that extracts the features and delivers the relevant features. This algorithm is used to standardise for main features and is related to the supervised learning algorithms. From this, the best features are extracted followed by the elimination of weakest features using the enhanced recursive elimination algorithm. From the selected features, an enhanced random forest classification is used to classify the results. The outcomes are executed to offer the best accuracy and thereby providing efficient prediction of effort estimation. Finally, the performance is measured with parameters such as Magnitude of Balanced Relative Error (MBRE), mean absolute residual, mean inverted balanced relative error, mean magnitude of error relative and mean magnitude of relative error. On comparing the existing methodologies, it is concluded that the proposed work offers better efficiency.This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Data types and amounts in human society are growing at an amazing speed, which is caused by emerging new services such as cloud computing and internet of things (IoT). As data has been a fundamental resource, research on big data has attracted much attention. An optimized cluster storage method for big data in IoT is proposed. First, weights of data blocks in each historical accessing period are calculated by temporal locality of data access, and the access frequencies of the data block in next period are predicted by the weights. Second, the hot spot of a data block is determined with a threshold that is calculated by previous data access. In this work, big data is divided into multiple segments based on semantic connectivity-based convolutional neural networks. Each segment will be stored in the different nodes by adapting the blockchain distributed-based local regenerative code technology called BCDLR. Experimental results demonstrate the efficiency of the proposed model in terms of packet delivery ratio, end-to-end delay, energy consumption, and throughput.

No abstract

No abstract