Programmers produce code clones when developing software. By copying and pasting code with or without modification, developers reuse existing code to improve programming productivity. However, code clones present challenges to software maintenance: they may require consistent application of the same or similar bug fixes or program changes to multiple code locations. To simplify the maintenance process, various tools have been proposed to automatically detect clones [1], [2], [3], [4], [5], [6]. Some tools tokenize source code, and then compare the sequence or frequency of tokens to reveal clones [1], [3], [4], [5]. Some other tools detect clones using tree-matching algorithms to compare the Abstract Syntax Trees (ASTs) of source code [2], [6]. In this paper, we present CCLEARNER, the first solely token-based clone detection approach leveraging deep learning. CCLEARNER extracts tokens from known method-level code clones and nonclones to train a classifier, and then uses the classifier to detect clones in a given codebase. To evaluate CCLEARNER, we reused BigCloneBench [7], an existing large benchmark of real clones. We used part of the benchmark for training and the other part for testing, and observed that CCLEARNER effectively detected clones. With the same data set, we conducted the first systematic comparison experiment between CCLEARNER and three popular clone detection tools. Compared with the approaches not using deep learning, CCLEARNER achieved competitive clone detection effectiveness with low time cost.

Researchers have developed various approaches for exploring the spatial information, temporal patterns, and Twitter content in topics of interest in order to generate a better understanding of human behavior; however, few investigations have integrated these three dimensions simultaneously. This study analyzes the content of tweets in order to conduct a spatiotemporal exploration of the main topics of interest in Kuwait in order to provide a deeper understanding of the topics people think about, when they think about them, and where they tweet about them. To this end, we collect, process, and analyze tweets from nearly 120 areas in Kuwait over a 10-month period. The study’s results indicate that religion, emotions, education, and public policy are the most popular topics of interest in Kuwait. Regarding the spatiotemporal analysis, people post more tweets regarding religion on Fridays, a holy day for Muslims in Kuwait. Moreover, people are more likely to tweet about policy and education on weekdays rather than weekends. In contrast, people tweet about emotional expressions more often on weekends. From the spatial perspectives, spatial clustering in topics occurs across the days of the week. The findings are applicable to further topic analysis and similar research in other countries.

Proximal communities in democratic societies comprise citizens, organizations, and governmental agencies, working together to identify and evaluate problems and opportunities in the public interest, build consensus around alternative approaches and solutions, and implement agreed upon policies. Communication and information sharing is critical to performing this collective work that involves not only face-to-face communication, but diverse media and technology whether offline (i.e., broadcast and print media) or online. We used topic modeling, social graphing and sentiment analysis to analyze information sharing behavior among individuals and organizations related to a geographic community and environs during municipal and state assembly elections in 2015. We investigate tweets, and Facebook posts and comments related to these elections as evidence for information sharing at the local level and/or of content relevant to the local community. Our findings suggest that the abundance of elections-relevant topics indicates that Twitter and FB were actively used for information sharing. The greater trust in local as opposed to non-local content and sources established in prior studies is consistent with our community level data in which sentiment expressed in our data is predominantly neutral. We argue that the greater trust in local as opposed to national sources of news, and in social media based on local social networks makes community-level groups and information sharing self-correcting and resilient, and thus, too small to fail.

To aid a variety of research studies, we propose TWIROLE, a hybrid model for role-related user classification on Twitter, which detects male-related, female-related, and brand-related (i.e., organization or institution) users. TWIROLE leverages features from tweet contents, user profiles, and profile images, and then applies our hybrid model to identify a user's role. To evaluate it, we used two existing large datasets about Twitter users, and conducted both intra- and inter-comparison experiments. TWIROLE outperforms existing methods and obtains more balanced results over the several roles. We also confirm that user names and profile images are good indicators for this task. Our research extends prior work that does not consider brand-related users, and is an aid to future evaluation efforts relative to investigations that rely upon self-labeled datasets.