Towards understanding the role of content-based and contextualized features in detecting abuse on Twitter.

This paper presents a novel approach for detecting abuse on Twitter. Abusive posts have become a major problem for social media platforms like Twitter. It is important to identify abuse to mitigate its potential harm. Many researchers have proposed methods to detect abuse on Twitter. However, most of the existing approaches for detecting abuse look only at the content of the abusive tweet in isolation and do not consider its contextual information, particularly the tweets posted before the abusive tweet. In this paper, we propose a new method for detecting abuse that uses contextual information from the tweets that precede and follow the abusive tweet. We hypothesize that this contextual information can be used to better understand the intent of the abusive tweet and to identify abuse that content-based methods would otherwise miss. We performed extensive experiments to identify the best combination of features and machine learning algorithms to detect abuse on Twitter. We test eight different machine learning classifiers on content- and context-based features for the experiments. The proposed method is compared with existing abuse detection methods and achieves an absolute improvement of around 7%. The best results are obtained by combining the content and context-based features. The highest accuracy of the proposed method is 86%, whereas the existing methods used for comparison have highest accuracy of 79.2%.

A Novel Image Encryption Technique Based on Mobius Transformation.

The nonlinear transformation concedes as S-box which is responsible for the certainty of contemporary block ciphers. Many kinds of S-boxes are planned by various authors in the literature. Construction of S-box with a powerful cryptographic analysis is the vital step in scheming block cipher. Through this paper, we give more powerful and worthy S-boxes and compare their characteristics with some previous S-boxes employed in cryptography. The algorithm program planned in this paper applies the action of projective general linear group PGL(2, GF(28)) on Galois field GF(28). The proposed S-boxes are constructed by using Mobius transformation and elements of Galois field. By using this approach, we will encrypt an image which is the preeminent application of S-boxes. These S-boxes offer a strong algebraic quality and powerful confusion capability. We have tested the strength of the proposed S-boxes by using different tests, BIC, SAC, DP, LP, and nonlinearity. Furthermore, we have applied these S-boxes in image encryption scheme. To check the strength of image encryption scheme, we have calculated contrast, entropy, correlation, energy, and homogeneity. The results assured that the proposed scheme is better. The advantage of this scheme is that we can secure our confidential image data during transmission.

Implementation of a Computerized Screening Inventory: Improved Usability Through Iterative Testing and Modification.

BACKGROUND: The administration of health screeners in a hospital setting has traditionally required (1) clinicians to ask questions and log answers, which can be time consuming and susceptible to error, or (2) patients to complete paper-and-pencil surveys, which require third-party entry of information into the electronic health record and can be vulnerable to error and misinterpretation. A highly promising method that avoids these limitations and bypasses third-party interpretation is direct entry via a computerized inventory. OBJECTIVE: To (1) computerize medical and behavioral health screening for use in general medical settings, (2) optimize patient acceptability and feasibility through iterative usability testing and modification cycles, and (3) examine how age relates to usability. METHODS: A computerized version of 15 screeners, including behavioral health screeners recommended by a National Institutes of Health Office of Behavioral and Social Sciences Research collaborative workgroup, was subjected to systematic usability testing and iterative modification. Consecutive adult, English-speaking patients seeking treatment in an urban emergency department were enrolled. Acceptability was defined as (1) the percentage of eligible patients who agreed to take the assessment (initiation rate) and (2) average satisfaction with the assessment (satisfaction rate). Feasibility was defined as the percentage of the screening items completed by those who initiated the assessment (completion rate). Chi-square tests, analyses of variance, and Pearson correlations were used to detect whether improvements in initiation, satisfaction, and completion rates were seen over time and to examine the relation between age and outcomes. RESULTS: Of 2157 eligible patients approached, 1280 agreed to complete the screening (initiation rate=59.34%). Statistically significant increases were observed over time in satisfaction (F3,1061=3.35, P=.019) and completion rates (F3,1276=25.44, P<.001). Younger age was associated with greater initiation (initiated, mean [SD], 46.6 [18.7] years; declined: 53.0 [19.5] years, t2,155=-7.6, P<.001), higher completion (r=-.20, P<.001), and stronger satisfaction (r=-.23, P<.001). CONCLUSIONS: In a rapid-paced emergency department with a heterogeneous patient population, 59.34% (1280/2157) of all eligible patients initiated the computerized screener with a completion rate reaching over 90%. Usability testing revealed several critical principles for maximizing usability of the computerized medical and behavioral health screeners used in this study. Further work is needed to identify usability issues pertaining to other screeners, racially and ethnically diverse patient groups, and different health care settings.

TrFAST: a tool to predict signaling pathway-specific transcription factor binding sites.

Recent advances in the development of high-throughput tools have significantly revolutionized our understanding of molecular mechanisms underlying normal and dysfunctional biological processes. Here we present a novel computational tool, transcription factor search and analysis tool (TrFAST), which was developed for the in silico analysis of transcription factor binding sites (TFBSs) of signaling pathway-specific TFs. TrFAST facilitates searching as well as comparative analysis of regulatory motifs through an exact pattern matching algorithm followed by the graphical representation of matched binding sites in multiple sequences up to 50kb in length. TrFAST is proficient in reducing the number of comparisons by the exact pattern matching strategy. In contrast to the pre-existing tools that find TFBS in a single sequence, TrFAST seeks out the desired pattern in multiple sequences simultaneously. It counts the GC content within the given multiple sequence data set and assembles the combinational details of consensus sequence(s) located at these regions, thereby generating a visual display based on the abundance of unique pattern. Comparative regulatory region analysis of multiple orthologous sequences simultaneously enhances the features of TrFAST and provides a significant insight into study of conservation of non-coding cis-regulatory elements. TrFAST is freely available at http://www.fi-pk.com/trfast.html.