Aleksander Wawer

Spiral of hatred: social effects in Internet auctions. Between informativity and emotion

An auction platform is a dynamic environment where a rich variety of social effects can be observed. Most of those effects remain unnoticed or even hidden to ordinary users. The in-depth studies of such effects should allow us to identify and understand the key factors influencing users’ behaviour. The material collected from the biggest Polish auction house has been analyzed. NLP algorithms were applied to extract sentiment-related content from collected comments and to measure informativity. Emotional distance between negative, neutral and positive comments has been calculated. The obtained results confirm the existence of the spiral-of-hatred effect but also indicate that much more complex patterns of mutual relations between sellers and buyers exist. The last section contains a several suggestions which can prove useful to improve trustworthiness of users’ reports and security of an auction platform in general.

Predicting webpage credibility using linguistic features

The article focuses on predicting trustworthiness from textual content of webpages. The recent work Olteanu et al. proposes a number of features (linguistic and social) to apply machine learning methods to recognize trust levels. We demonstrate that this approach can be substantially improved in two ways: by applying machine learning methods to vectors computed, using psychosocial and psycholinguistic features and in a high-dimensional bag-of-words paradigm of word occurrences. Following Olteanu et al., we test the methods in two classification settings, as a 2-class and 3-class scenario, and in a regression setting. In the 3-class scenario, the features compiled by Olteanu et al. achieve weighted precision of 0.63, while the methods proposed in our paper raise it to 0.66 and 0.70. We also examine coefficients of the models in order to discover words associated with low and high trust.

Temporal, Cultural and Thematic Aspects of Web Credibility

Is trust to web pages related to nation-level factors? Do trust levels change in time and how? What categories (topics) of pages tend to be evaluated as not trustworthy, and what categories of pages tend to be trustworthy? What could be the reasons of such evaluations? The goal of this paper is to answer these questions using large scale data of trustworthiness of web pages, two sets of websites, Wikipedia and an international survey.

Expanding Opinion Attribute Lexicons

The article focuses on acquiring new vocabulary used to express opinion attributes. We apply two automated expansion techniques to a manually annotated corpus of attribute-level opinions. The first method extracts opinion attribute words using patterns. It has been augmented by the second, wordnet and similarity-based expansion procedure. We examine the types of errors and shortcomings of both methods and end up proposing two hybrid, machine learning approaches that utilise all the available information: rules, lexical and distributional. One of them proves highly successful.

The Scent of Deception: Recognizing Fake Perfume Reviews in Polish

Many opinion aggregation websites are injected by well- formed, fake reviews. Previous work showed that such spam is very hard to identify by a non-expert human reader and therefore, automated methods are needed to identify deceptive content. To this day, there has been no study for the Polish language and our main goal is to fill that gap. We present a corpus of fake and true reviews in Polish and describe experiments on automated opinion spam detection. Our approach turns out to be highly successful, but future systematic studies are needed to confirm the nature of our findings.

Application of evolutionary computing to conformational analysis

In this paper we report new results concerning use of genetic algorithms in conformational analysis, field of pharmacy related to discovery and design of new drugs. The goal is to find the optimal spatial configuration of a molecule, which corresponds to finding its energy minimum by rotation of torsion angles. A number of experimental results obtained with different evolutionary mechanisms and parameters and further evaluate them on a real-life example of vitamine E is presented.

Do We Need Word Sense Disambiguation for LCM Tagging

Observing the current state of natural language processing, especially in the Polish language, one notices that sense-level dictionaries are becoming increasingly popular. For instance, the largest manually annotated sentiment dictionary for Polish is now based on plWordNet (the Polish WordNet) [13], also the Polish Linguistic Category Model (LCM-PL) [10] dictionary has its significant part annotated on sense level. Our paper addresses the important question: what is the influence of word sense disambiguation in real-world scenarios and how it compares to the simpler baseline of labeling using just the tag of the most frequent sense. We evaluate both approaches on data sets compiled for studies on fake opinion detection and predicting levels of self-esteem in the area of social psychology. Our conclusion is that the baseline method vastly outperforms its competitor.

Collective Memory in Poland: A Reflection in Street Names

Our article starts with an observation that street names fall into two general types: generic and historically inspired. We analyse street names distributions (of the second type) as a window to nation-level collective memory in Poland. The process of selecting street names is determined socially, as the selections reflect the symbols considered important to the nation-level society, but has strong historical motivations and determinants. In the article, we seek for these relationships in the available data sources. We use Wikipedia articles to match street names with their textual descriptions and assign them to the time points. We then apply selected text mining and statistical techniques to reach quantitative conclusions. We also present a case study: the geographical distribution of two particular street names in Poland to demonstrate the binding between history and political orientation of regions.

Collaborative Problem Solving in Emergency Situations: Lessons Learned from a Rescue Mission

The focus of this paper is on time-constrained collaborative problem solving using common web-based communication systems. The rescue action of a missing Polish kite surfer conducted with help of the kiteforum.pl community created a unique opportunity to investigate how people organize such actions in the Internet and how they collaborate under the enormous pressure of time. Quantitative and qualitative analysis of phenomena that occurred during the people’s interaction, has been done with help of natural language processing techniques. A list of recommendations for the designers of collaborative problem solving systems and people involved in such action has been proposed.

Spiral of Hatred: Social Effects in Buyer-Seller Cross-Comments Left on Internet Auctions

An auction platform is a dynamic environment where a rich variety of social effects can be observed. Most of those effects remain unnoticed or even hidden to ordinary users. The in-depth studies of such effects should allow us to identify and understand the key factors influencing users’ behaviour. The material collected from the biggest Polish auction house has been analyzed. NLP algorithms were applied to extract sentiment-related content from collected comments. Emotional distance between negative, neutral and positive comments has been calculated. The obtained results confirm the existence of the spiral-of-hatred effect but also indicate that much more complex patterns of mutual relations between sellers and buyers exist. The last section contains a several suggestions which can prove useful to improve trustworthiness of users’ reports and security of an auction platform in general.