The majority of documented phishing attacks have been carried by email, yet few studies have measured the impact of email headers on the predictive accuracy of machine learning techniques in detecting email phishing attacks. Research has shown that the inclusion of a limited subset of email headers as features in training machine learning algorithms to detect phishing attack did increase the predictive accuracy of these learning algorithms. The same research also recommended further investigation of the impact of including an expanded set of email headers on the predictive accuracy of machine learning algorithms. In addition, research has shown that the cost of misclassifying legitimate emails as phishing attacks--false positives--was far higher than that of misclassifying phishing emails as legitimate--false negatives, while the opposite was true in the case of fraud detection. Consequently, they recommended that cost sensitive measures be taken in order to further improve the weighted predictive accuracy of machine learning algorithms. Motivated by the potentially high impact of the inclusion of email headers on the predictive accuracy of machines learning algorithms and the significance of enabling cost-sensitive measures as part of the learning process, the goal of this research was to quantify the impact of including an extended set of email headers and to investigate the impact of imposing penalty as part of the learning process on the number of false positives. It was believed that if email headers were included and cost-sensitive measures were taken as part of the learning process, than the overall weighted, predictive accuracy of the machine learning algorithm would be improved. The results showed that adding email headers as features did improve the overall predictive accuracy of machine learning algorithms and that cost-sensitive measure taken as part of the learning process did result in lower false positives.

京都大学 / 0048 / 新制・課程博士 / 博士(情報学) / 甲第21975号 / 情博第703号 / 新制||情||121(附属図書館) / 京都大学大学院情報学研究科通信情報システム専攻 / (主査)教授 佐藤 高史, 教授 小野寺 秀俊, 教授 岡部 寿男 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DFAM

Im Rahmen dieser Studienarbeit wird untersucht, welche Lösungen es momentan zur Filterung von Email mit unerwünschten Schadensfunktionen auf Mailservern gibt. Dabei werden sowohl offene als auch proprietäre Lösungen betrachtet und die momentanen Randbedingungen der TU Chemnitz beachtet.

Email is an essential form of communication for organizations. Nevertheless, with so much popularity came many challenges. These emails usually carry sensitive data that might cause significant harm if they get compromised. Besides, spam and phishing emails that continually reach the employees’ inbox masquerading as a trusted entity due to the lack of authentication mechanisms are also considered a significant threat for organizations today. Such threats are phishing using email domain forgery attack, redirecting emails to a mail server that is under the attacker’s control, and connection eavesdropping. The research aimed to investigate the vulnerability of approximately 2000 organizations within Sweden against those attacks. Toward that end, the quantity and quality of the following email security mechanisms SPF, DKIM, DMARC, STARTTLS, DNSSEC, and DANE were examined through a case study. Also, the adoption of these mechanisms was investigated, whether it varies based on different factors such as organization size, sector, and location. The research findings indicated that the average adoption rate by the tested organizations was approximately 50%. Furthermore, the result demonstrated that there were no differences in the adopted mechanisms based on the studied factors that the results were quite similar among the tested groups. It concluded that there is a lack of protection mechanisms, which made the majority of the tested organizations vulnerable to different types of email attacks.

Classifying emails into distinct labels can have a great impact on customer support. By using machine learning to label emails the system can set up queues containing emails of a specific category. This enables support personnel to handle request quicker and more easily by selecting a queue that match their expertise. This study aims to improve the manually defined rule based algorithm, currently implemented at a large telecom company, by using machine learning. The proposed model should have higher F1-score and classification rate. Integrating or migrating from a manually defined rule based model to a machine learning model should also reduce the administrative and maintenance work. It should also make the model more flexible. By using the frameworks, TensorFlow, Scikit-learn and Gensim, the authors conduct five experiments to test the performance of several common machine learning algorithms, text-representations, word embeddings and how they work together. In this article a web based interface were implemented which can classify emails into 33 different labels with 0.91 F1-score using a Long Short Term Memory network. The authors conclude that Long Short Term Memory networks outperform other non-sequential models such as Support Vector Machines and ADABoost when predicting labels for emails.

Denna rapport utvärderar olika metoder för att klassificera och kategorisera email. Mångamail anländer hos människors inkorg varje dag. När tiden går och antalet email ökar blir detsvårare att hitta specifika email. På HDAB arbetar de som konsulter och vill dela upp email iolika mappar beroende på vilket projekt det tillhör. Idag fungerar det genom ett ord-regelbaseratsystem som sorterar email I olika mappar med en precision på cirka 85%. HDAB villta reda på om det går att använda maskininlärning för det nuvarande systemet. Denna rapportpresenterar fyra maskininlärningsalgorimer, beslutsträd, random forest beslutsträd, k-nearestneighbor och naive bayes, som användas för att utvärdera om det är möjligt att kategoriseraemailen.Datan som används till rapporten kommer från HDABs mailserver och är redan kategoriseradtill rätt kaegori. / This report will evaluate methods for classifying e-mails into different categories. A lot ofemails are received in peoples inboxes every day. When the time passes and the amount ofemails increases the ability to find specific emails gets harder. At HDAB they are workingwith consulting and want to separate different emails from different project into separate folders.This is achieved today by using a word based rule system that sorts emails into differentfolders and has a precision about 85%. HDAB wants to know if it is possible to use machinelearning to automatically sort the emails into different folders instead of the current solution.This report presents four machine learning algorithms, decision tree, random forest decisiontree, k-nearest neighbor and naive bayes, which are being used for evaluation of the possibilityto categorize the emails.The data used for the report will be data gathered from HDAB’s mail server and are alreadypre-labeled into their respectively categories.

The aim of the thesis called Some specifics of communication in civil engineering has been to carry out research in construction companies and to find out to what extent they pay attention to corporate image in various respects – if (and eventually how) they make use of the Internet and how they present themselves to the public. On the other hand, part of the research focused on what emphasis do these companies put on internal communication with their employees. To prove or displace the initial hypotheses, several analyses have been made and research in 14 companies of different size carried out. The results have shown that construction companies pay attention to internal communication, however, they do not place much emphasis on presenting themselves to public. This way they underestimate the importance of positive corporate image, particularly in the media of the Internet.

Background: As personalization has become a common CRM strategy for companies to create valuable relationships with customers, users are receiving an increased amount of personalized communication, further research is needed on the influence of content personalization in specific channels, to improve customer engagement.  Purpose: This paper seeks to analyse the influence of eCRM content personalization strategies on the consumer engagement performance of email marketing campaigns building upon existing knowledge about the benefits and opportunities of personalization strategies.  Method: The selected method is quantitative, using A/B testing, based on the comparison of different variations on two identical segments where the only differentiated element is the one being evaluated. Two controlled experiments evaluating subject line and image personalization are performed evaluating four metric of email marketing performance: open rate, click-through rate, conversion rate and unsubscribe rate.  Conclusion: The results of the two controlled experiments performed for this research, subject line personalization and image personalization, complement past literature on content personalization strategies (Bertrand et al., 2010; Carlota Rocha et al., 2023; Munz et al., 2020; Sahni et al., 2018) by demonstrating the statistically significant positive influence of content personalization on email marketing performance and reinforcing the importance of familiarity with the brand or product on reducing the probability of bad outcomes.

The study introduces the concept of attitudes and its fundamentals, as well as how it revolvesaround the given attitude object, in this instance brands. Furthermore, it introduces thephenomena of email marketing and how it has evolved its consensual form that is mostcommon today, referred to as PEM (Permission-based email marketing) in this study.Conclusively explaining the important influences between these two as well as the researchgap in this given context. The final purpose of the paper is to explain the impact thatpermission-based email marketing has on brand attitude. The methodology of the study itselfutilized a cross-sectional approach with a self-completion questionnaire via a conveniencesampling through a web-based survey. The final sample consisted of 108 respondents thatwere deemed acceptable. The study concluded that entertaining content and informativecontent seemingly affects brand attitude positively, whereas the impact of frequency couldnot definitively be explained. Finally theoretical and practical implications of the results arediscussed, as well as limitations of the study and recommended future research.

Under läsåret 2017/2018 var det över 700 000 registrerade studenter på gymnasie och högskolenivå i Sverige. En fördel med att studera är att det finns en marknad för erbjudanden med reducerade priser riktade till just studenter. Många av dessa erbjudanden gäller köp online och ett vanligt sätt att marknadsföra dem är genom email. Denna studie undersöker hur stor stor effekt marknadsföring via email har på studenters digitala konsumtion. Utöver att undersöka den generella effekten på köpmängden studeras även vilken påverkan faktorerna produktkategori och tidpunkt för email-utskicken har. Dessutom undersöks studenters attityd gentemot email som marknadsföringskanal. Syftet med studien är att öka förståelsen för studenters köpbeteende och därmed få underlag till att kunna optimera marknadsföring via email. Därav har det undersökts hur studenter reagerar på email-utskick genom att jämföra antalet köp som skett under dagarna innan och efter utskicken. Därefter har det genomförts en statistisk analys med hypotesprövning av de observerade förändringarna. Resultatet visar att email-marknadsföring leder till en signifikant konsumtionsökning då ett intervall på tre dagar runt utskicksdatumet studeras. Denna ökning sker för de båda studerade produktkategorierna och tidpunkterna. Dock visar resultatet att olika produktkategorier har olika stor påverkan på email-kampanjernas effekt. Studien kunde inte påvisa någon skillnad i påverkan för olika tidpunkter, något som kan bero på för liten datamängd. Slutligen visar resultatet att majoriteten av studenterna kontrollerar sin email-inkorg dagligen, men att många slänger email med erbjudanden direkt i papperskorgen. Slutsatsen är att email är en effektiv marknadsföringskanal för att nå studenter, men att det finns behov av optimering för att utnyttja dess potential bättre. / During the school year 2017/2018 there were over 700,000 registered students at high school or university level in Sweden. A benefit of being a student is that there is a market for discounts and offers aimed specifically at students. Many of these offers apply online and a common way to promote them is through email. This study examines the impact of email marketing on students digital consumption. In addition to examining the general effect, the impact of the factors product category and timing of the emails is studied. Furthermore, students’ attitude towards email as a marketing channel has been investigated. The purpose of this study is to increase the understanding of students’ buying behavior in order to optimize email marketing. Therefore, students’ reaction to email marketing has been investigated by analyzing data and comparing the number of purchases made during the days before and after the mailings. Subsequently, a statistical analysis has been carried out with hypothesis testing of the observed changes. The result shows that email marketing leads to a significant increase in consumption when an interval of three days around the email date is studied. This increase occurs for both product categories and timings examined. However, the result shows that different product categories have different effects on the impact of the email. The study did not give enough documentation to be able to say if some timing is more effective than others. Moreover, this study shows that the majority of students check their mailbox daily, but that many throw emails containing offers directly in the trash. The conclusion is that email is an effective marketing channel for reaching students, but that optimization is needed to make better use of its potential.