Ali Mert Ertugrul

An exploratory study on role-based collaborative business process modeling approaches

Role-based business process modeling allows roles to focus on modeling their own parts of work and requires them to negotiate with each other in order to form a consistent and integrated process. Due to the collaborative nature of role-based modeling, negotiations among roles have a crucial impact on the overall process. Role-based modeling approaches differ from each other based on the differences in their negotiation way. In this study we analyzed three current role-based collaborative business process modeling approaches in the literature, namely ex-ante, ex-post and ongoing communication negotiation approaches and presented their benefits and the drawbacks by conducting an exploratory research. By modeling a sample process using current approaches with S-BPM notation, we identified the weaknesses and strengths of each approach. We also discuss possible solutions to these drawbacks and suggested an ideal role-based collaborative business process modeling approach that combines the advantages of these approaches with our proposed solutions.

Word Embedding Based Event Detection on Social Media

Event detection from social media messages is conventionally based on clustering the message contents. The most basic approach is representing messages in terms of term vectors that are constructed through traditional natural language processing (NLP) methods and then assigning weights to terms generally based on frequency. In this study, we use neural feature extraction approach and explore the performance of event detection under the use of word embeddings. Using a corpus of a set of tweets, message terms are embedded to continuous space. Message contents that are represented as vectors of word embedding are grouped by using hierarchical clustering. The technique is applied on a set of Twitter messages posted in Turkish. Experimental results show that automatically extracted features detect the contextual similarities between tweets better than traditional feature extraction with term frequency - inverse document frequency (TF-IDF) based term vectors.

Effect of using regression in sentiment analysis

In this study, the effect of using regression on sentiment classification of Twitter data was analyzed. In other words, whether the strength of sentiment better discriminates the classes or not. Since our dataset includes class confidence scores rather than discrete class labels, regression analysis was employed on each class separately. Then, each tweet was assigned the class whose estimated confidence score is maximum among others after regression. The feature set used includes unigrams, POS tags, emoticons, sentiments of words and POS tags of sentiments. The results of experiments indicate that using classification on discrete class labels perform much better than using regression on continuous confidence scores.

Does the Strength of Sentiment Matter? A Regression Based Approach on Turkish Social Media

Social media posts are usually informal and short in length. They may not always express their sentiment clearly. Therefore, multiple raters may assign different sentiments to a tweet. Instead of employing majority voting which ignores the strength of sentiments, the annotation can be enriched with a confidence score assigned for each sentiment. In this study, we analyze the effect of using regression on confidence scores in sentiment analysis using Turkish tweets. We extract hand-crafted features including lexical features, emoticons and sentiment scores. We also employ word embedding of tweets for regression and classification. Our findings reveal that employing regression on confidence scores slightly improves sentiment classification accuracy. Moreover, combining word embedding with hand-crafted features reduces the feature dimensionality and outperforms alternative feature combinations.

A Method for Modeling Business Processes in a Role-based and Decentralized Way

Role-based and decentralized process modeling allows actors to focus on modeling their own role behaviors and requires them to negotiate to form a consistent and integrated process model. Negotiations among the actors have a crucial impact on modeling activity of overall process since decentralized process modeling has a collaborative nature. Based on the time spent for negotiation among the actors, process modeling approaches vary. In this study, we propose a role-based and decentralized process modeling method, called ROADMap that enables process stakeholders to model their own works in a bottom up manner. In order to observe the implications of the proposed method, we conducted a case study. Case study results indicate that ROADMap method is a successful role-based and decentralized process modeling method for process stakeholders.

Effect of Using Regression on Class Confidence Scores in Sentiment Analysis of Twitter Data

In this study, we aim to test our hypothesis that confidence scores of sentiment values of tweets aid in classification of sentiment. We used several feature sets consisting of lexical features, emoticons, features based on sentiment scores and combination of lexical and sentiment features. Since our dataset includes confidence scores of real numbers in [0-1] range, we employ regression analysis on each class of sentiments. We determine the class label of a tweet by looking at the maximum of the confidence scores assigned to it by these regressors. We test the results against classification results obtained by converting the confidence scores into discrete labels. Thus, the strength of sentiment is ignored. Our expectation was that taking the strength of sentiment into consideration would improve the classification results. Contrary to our expectations, our results indicate that using classification on discrete class labels and ignoring sentiment strength perform similar to using regression on continuous confidence scores.

RemindMe: An Enhanced Mobile Location-Based Reminder Application

In this study, a location based reminder application RemindMe, enhanced with various location tagging options using social networking APIs is proposed. Main purpose of this application is to allow users to create reminders based on the location besides time and to notify users with those reminders automatically. In terms of ease of use, a hybrid structure consisting of various components is formed for location tagging. First of all, the user tags the locations using the applications such as Google Maps or Foursquare or via the embedded sensors of the Android device. Then, he creates reminders for the tagged locations and when he gets close to this location, the system notifies the user. Our application is separated from similar applications with its enhanced location tagging feature. Moreover, by consisting of various services, it is open to innovations on the way to become a social reminder application. The usability test results indicate that RemindMe is an effective location based reminder application.

The Effect of Highlighting Error Categories in FSM Training on the Accuracy of Measurement

As various software management activities including cost estimation and project control are conducted based on the software size measurement, achieving high accuracy in functional size measurement (FSM) is critical. Several studies examined the relation between FSM training and improvement in the accuracy of FSM. However, those studies propose comprehensive frameworks and approaches that require fundamental changes in the training content. In this study, we analyzed the effect of highlighting error categories during training by extracting the errors throughout four years. We showed that, highlighting the frequent error categories during the same training period without a fundamental change in the content would significantly decrease the error rate. The results of the research we conducted are promising about the improvement of measurement accuracy.