The impact of deep learning on document classification using semantically rich representations

This paper presents a semantically rich document representation model for automatically classifying financial documents into predefined categories utilizing deep learning. The model architecture consists of two main modules including document representation and document classification. In the first module, a document is enriched with semantics using background knowledge provided by an ontology and through the acquisition of its relevant terminology. Acquisition of terminology integrated to the ontology extends the capabilities of semantically rich document representations with an in depth-coverage of concepts, thereby capturing the whole conceptualization involved in documents. Semantically rich representations obtained from the first module will serve as input to the document classification module which aims at finding the most appropriate category for that document through deep learning. Three different deep learning networks each belonging to a different category of machine learning techniques for ontological document classification using a real-life ontology are used.Multiple simulations are carried out with various deep neural networks configurations, and our findings reveal that a three hidden layer feedforward network with 1024 neurons obtain the highest document classification performance on the INFUSE dataset. The performance in terms of F1 score is further increased by almost five percentage points to 78.10% for the same network configuration when the relevant terminology integrated to the ontology is applied to enrich document representation. Furthermore, we conducted a comparative performance evaluation using various state-of-the-art document representation approaches and classification techniques including shallow and conventional machine learning classifiers.     

Effects of Conceptual Change Texts and Laboratory Experiments on Fourth Grade Students’ Understanding of Matter and Change Concepts

The purpose of this study was to investigate whether conceptual change texts and laboratory experiments are effective in overcoming misconceptions and whether the concepts were acquired permanently when these methods were utilized. In this study, we addressed some topics from the “Matter and Change” unit in science and technology class of elementary 4th grade. Students from three classes of an elementary school participated in the study (N = 104). Students’ misconceptions were determined by administering the “Matter Concept Test” before, immediately after and 13 weeks after the instructional period. The results of the study showed that both conceptual change texts and experiment method were more successful than traditional instruction in overcoming the misconceptions and acquiring permanent knowledge. However, there was not a significant difference between these two alternative approaches in terms of reducing the misconceptions.     

Misconceptions of Turkish Pre-Service Teachers about Force and Motion

The purpose of this study was to diagnose the misconceptions held by pre-service physics teachers about force and motion. The secondary aim of the study was to detect whether misconceptions vary according to gender, educational level, and culture. The study was conducted with 79 student-teachers attending to one of the largest faculties of education in Turkey. Force Concept Inventory (FCI) was used to diagnose student-teachers’ misconceptions. FCI is a conceptual test consisting of 29 multiple choice items. Each wrong choice for each question reflects a specific misconception about the force and motion concepts. Data from the study was analyzed by using frequencies, t-test, and ANOVA for making comparisons according to gender and years of education. Results of the study showed that student-teachers of physics hold very strong misconceptions about impetus and active force. No significant differences were found between male and female students’ scores on the concept test. The results also showed that misconceptions about force and motion decreased through the years of education. However, they did not disappear completely. Findings of the study are very similar to the other research findings conducted on the subject in other countries. Student-teachers’ conceptions about Newton’s Third Law, on the other hand, were significantly better than those observed in other research done in other countries such as the US and Finland.     

Teaching about homosexualities to Nigerian university students: a report from the field

Nigeria’s diverse cultures, religions and political parties appear to be unified by a strong taboo against homosexuality and gay rights. This has affected academic research, HIV/AIDS programmes, and sexuality education, all which commonly show evidence of heterosexism, self-censorship and even explicit condemnations of homosexuality. Yet a dissident discourse in Nigeria, as well as research from elsewhere in the region, suggests that this appearance of unity may belie greater openness to the issues than assumed. Indeed, research shows that (1) many African societies are traditionally more accommodating toward non-normative sexualities than contemporary nationalist or cultural claims would allow, and (2) secretive ‘bisexuality’ is more common in practice (and tacitly acknowledged) than previously understood. Is it possible then that the presumption of homophobia and the fear of backlash that has clearly contributed to heterosexism and self-censorship in scholarship around homosexualities in Nigeria are exaggerated? Is it possible that Nigerians may be more open to consideration of scientific evidence and international best practices around sexual diversity, rights, and health than is commonly assumed in the literature? A trial intervention at a small state university in a predominantly rural area of Nigeria tested these questions by introducing wide-ranging, frank and non-judgemental (science-based) discussions of same-sex sexuality in several classes. Analysis of the students’ feedback finds that stigmatising attitudes toward homosexuality were indeed present among the students. However, there was also a high degree of curiosity, awareness of the existence of secretive homosexualities in Nigeria, desire for education, and confidence that traditional cultures and Nigerian democracy could accommodate individual freedom and sexual rights. The conclusion is that well-prepared researchers and educators could be less anxious and self-censoring around the topic of homosexuality than prevails at present. Careful attention would need to be paid to local sensibilities, but sexuality and HIV education programmes could probably be brought closer into line with world guidelines on best practices and comprehensive approaches to human sexuality education and sexual health.     

Effects of specimen haemolysis on complete blood count results by Abbott Alinity hq System

IntroductionThe current study aimed to assess the interference of in vitro haemolysis on complete blood count (CBC) using Abbott Alinity hq system, and to determine which haemolysis levels affect the reliability of sample results.Materials and methodsBlood samples obtained from 25 volunteers in K3-EDTA tubes were divided into four aliquots. The first aliquot was not subjected to any intervention. The second, third and fourth aliquots were passed through a fine needle 2, 4 and 6 times, respectively. Complete blood count was performed by multi-angle polarized scatter separation technology and haemolysis index (HI) was assessed from the plasma samples separated by centrifugation. Five groups were formed according to the HI values. The percentage biases between the results of non-haemolysed and haemolysed groups were compared with the desirable bias limits from The European Federation of Clinical Chemistry and Laboratory Medicine database and reference change values (RCVs).ResultsIn groups 1 to 4, the effects of haemolysis on CBC parameters were acceptable comparing to the analytical bias except for lymphocytes (7.26%-7.42%), MCH (2.59%), and MCHC (0.47%-2.81%). Results of group 5 (gross haemolysis) showed decreases in HCT(- 4.56%), RBC (- 4.07%) count and increase in lymphocyte (11.60%) count higher than the analytical performance specifications. Moreover, variations in MCH (4.65%) and MCHC (5.24%) were exceeding the RCVs.ConclusionsGross haemolysis (haemoglobin concentration > 10 g/L) is likely to produce unreliable CBC results on non-pathological samples. Further studies including pathological specimens are needed.