With the exponential growth of Internet and the various forms of social media, the number of people using such platforms to share their views and experiences is also greatly increased. The reviews posted by people regarding a service or entity serve as valuable sources for decision making. However, acquiring valuable insights from the abundantly available unstructured information is not so straightforward. There is a need for more effective methodologies to carry out aspect-based sentiment analysis in a fine-grained manner. The proposed work performs joint aspect-opinion extraction and sentiment orientation detection (JAESOD) using an integrated approach of complex dependency rule-based extraction and proper treatment of various forms of adjectives that contribute to effective extraction of subjective sentiment-bearing terms. The objective of the proposed work is three-fold: to extract aspect-sentiment word pairs, to detect implicit aspects and to perform sentiment orientation detection at the aspect level along with sentiment scoring. In this paper, we propose a novel methodology to classify the sentiment orientation on a fine-tuned nine-point scale in contrast to the existing three to five point scale found in the literature works. With respect to the task of sentiment detection, the goal of the proposed work is to achieve optimality in sentiment classification by giving due weightage to the modifies or intensifiers based on their degree of intensification and fair treatment of crucial factors such as double intensification, extended forms of adjectives, negations, sentiments expressed in numerical forms that play a vital role in fine tuning the sentiment orientation detection.

Mobile cloud computing is used to define and determine computing services with a structure model. The data and resource of any service will be retrieved from cloud computing through internet service, some tools, and user interface (web-based or application). Mobile Cloud Computing (MCC) is a hybrid of cloud computing and mobile computing. Multimedia Information is the core of Mobile Cloud information because of the sizable information of multimedia particularly video streaming. Mobile Cloud mostly handles and processes that information. MCC is one of the business expressions with the real environment in the IT world. The concept of the MCC is still in the beginner stage of advancement. So, the handle of the innovation in a careful way especially in the bearing of future research should be provide. In this paper, an algorithm is throttled load balancing for mobile clouds has been presented within an example of Multimedia information. The results has shown that the load balancing of cloud computing environment. In this scenario, load balancing techniques in mobile cloud computing can be employed and can successfully manage time through the cloud.

The purpose of this paper is to provide an overview of the major studies that have addressed sentiment analysis in Arabic using deep learning. Based on the findings of the review that we have conducted, it has been revealed that various models have been used. These include convolutional neural networks (CNNs), long short-term memory (LSTM), bidirectional recurrent neural networks (Bi-RNNs), deep neural networks (DNNs), gated recurrent units (GRUs), as well as attention-based neural networks. These models have explored sentiment analysis in this morphologically complex language at various levels, namely sentence, document, and aspect levels. At sentence level, it has been found out that LSTM approaches outperform other deep learning models such as CNN on most datasets. At document level, CNN has been reported to give the best results followed by LSTM. However, at aspect level, it has been demonstrated that LSTM is the most commonly used model by researchers. The findings of the present study have also proven that deep learning models are more effective on larger datasets compared to basic machine learning classifiers such as SVM, and NB and that word embedding representations are much more efficient than classical techniques.

An important concept of machine learning is Feature selection. It is the process of reducing the inputs for processing and analysis, or of finding the most meaningful inputs. Feature engineering extracts useful information or features from existing data. It is a major area of innovative practical outcome and established to answer the challenges due to data of increasingly high dimensionality. The tangible benefits comprises of constructing simpler and more intelligible models, improving data mining performance, and facilitates data preparation task in a feasible approach. Initially it has been briefed about the key elements of feature selection, analysed about the progress with the growth of data mining techniques and the role of feature selection in bio medicine.

T-spherical fuzzy set is one of the most generalized concept in fuzzy set theory which can deal with membership, neutral and non-membership degrees simultaneously. In this paper, we proposed the concept of linguistic interval-valued T-spherical fuzzy set (LIVt-SFS). In it, the membership, neutral and non-membership degrees are characterized by interval valued linguistic terms for better dealing with the uncertainty, ambiguity in group decision-making processes. To explore the study, we developed some basic operations, score and accuracy functions to compare two or more linguistic interval-valued T-spherical fuzzy numbers (LIVt-SFNs) and their desirable properties are also discussed. Based on these operations two aggregation operators such as LIVt-SF weighted averaging and LIVt-SF weighted geometric operators are proposed. Moreover, technique for order of preference by similarity to ideal solution (TOPSIS) method is developed to solve multi-attribute group decision-making (MAGDM) under LIVt-SFS information. Furthermore, a MAGDM technique is proposed based on aggregation operators and TOPSIS method, then practical example is given in order to explain the proposed method. Finally, a comparison analysis is undertaken to demonstrate the efficacy and applicability of the proposed method.

Extractive document summarization (EDS) is usually seen as a sequence labelling task, which extracts sentences from a document one by one to form a summary. However, extracting sentences separately ignores the relationship between the sentences and documents. One solution is to use sentence position information to enhance sentence representation, but this will cause the sentence-leading bias problem, especially in news datasets. In this paper, we propose a novel sentence centrality for the EDS task to address these two problems. The sentence centrality is based on directed graphs, while reflecting the sentence-document relationship, it also reflects the sentence position information in the document. We implicitly strengthen the relevance of sentences and documents by using sentence centrality to enhance sentence representation. Notably, we replaced the sentence position information with sentence centrality to reduce sentence-leading bias without causing model performance degradation. Experiments on the CNN/Daily Mail dataset showed that EDS models with sentence centrality significantly improved compared with baseline models.

The Distributed Python Software Transactional Memory (DPSTM) is a fault-tolerant, deterministic, and Python-based distributed software transactional memory, which targets IoT embedded systems at the edge of the internet, such as smart homes, cars, etc. This paper presents DPSTM’s formal verification using the process algebra communicating sequential processes and the accompanying model checker PAT. The formal verification is conducted through the modelling and verification phases. In the modelling phase, the relevant DPSTM functions are modelled as the corresponding communicating sequential processes, whereas in the verification phase, the appropriate test systems are created and their safety and liveness properties are automatically verified by PAT. The test systems were designed with the goal to formally verify correctness of the three key DPSTM’s features: (i) the distributed master-slave finite state machine governing a pair of transaction coordinators working in the master-slave mode, (ii) the data replication protocol, and (iii) the procedure for fixing broken service after the master transaction coordinator’s crash. During verification, altogether 36 assertions were automatically checked by PAT and found to be valid.