Classifying Emotions in Film Music—A Deep Learning Approach

The paper presents an application for automatically classifying emotions in film music. A model of emotions is proposed, which is also associated with colors. The model created has nine emotional states, to which colors are assigned according to the color theory in film. Subjective tests are carried out to check the correctness of the assumptions behind the adopted emotion model. For that purpose, a statistical analysis of the subjective test results is performed. The application employs a deep convolutional neural network (CNN), which classifies emotions based on 30 s excerpts of music works presented to the CNN input using mel-spectrograms. Examples of classification results of the selected neural networks used to create the system are shown.

Peningkatan Kemampuan Menulis Surat Lamaran Pekerjaan Melalui Model Pembelajaran Think-Pair-Share

The purpose of the study was to find out how the ability to write job application letters for class XII students of SMK Negeri 2 Depok City in 2019-2020 and to find out whether cooperative learning with the Think-Pair-Share model could improve the ability to write cover letters for class XII students of SMK Negeri 2 Depok City. This study uses classroom action research (action research) in two rounds (cycles). Each round consists of four stages, namely: planning, activities and observations, reflection, and revision. The target of this research is Class XII Computer Network Engineering (TKJ) SMK Negeri 2 Depok City in 2019-2020. The data obtained in the form of subjective test results and observation sheets of teaching and learning activities. The results show, class XII students of TKJ SMK Negeri 2 Depok City in 2019-2020 are able to write job applications. This can be seen from the average score of students' ability to write job application letters and student learning completeness in cycle I and cycle II has increased. The average score of students' ability to write job application letters in the first cycle was 72.77 and learning completeness reached 64.70%. The average score of students' ability to write job application letters in cycle II was 78.38 and learning completeness reached 85.29%. The application of the Think-Pair-Share model of cooperative learning has a positive influence on Teaching and Learning Activities (KBM). Think-Pair-Share model of cooperative learning can improve student learning outcomes. This is indicated by the average student answers stating that students are interested and interested in the Think-Pair-Share cooperative learning method so that they become motivated to learn. Keywords: Writing Ability, Cooperative Learning, Think-Pair-Share Learning Model

TEST ANXIETY AS A PREDICTOR OF ACADEMIC DISHONESTY AMONG UNDERGRADUATES IN TERTIARY INSTITUTIONS IN KOGI STATE

Academic dishonesty is a critical educational concern across all educational levels in the world. The trend is pervasive among higher education students and undermines human and national developments. This study examined academic dishonesty among students in tertiary institutions in Kogi State, Nigeria, based on their subjective test anxiety. One hundred and thirty-seven undergraduates participated in the study. The students were approached between July and September 2021, and they completed self-report measures of the Cognitive Test Anxiety Scale and the Academic Dishonesty Scale. The data were analyzed using a simple linear regression model. The finding revealed that test anxiety accounted for 56.2% of the variation in academic dishonesty. Our result supports literature indicating a link between test anxiety and academic dishonesty among higher education students. We recommend that a simulation exercise is vital in preparing students for assessments.

Sound quality evaluation for luxury refrigerator door closing sound

In this study, we carried out subjective evaluation tests employing 19 refrigerator door closing sounds to quantify the luxury feeling. By applying factor analysis to the subjective evaluation results, the sound quality of the refrigerator door closing sound was found to be expressed by the following two factors: overall loudness and the pitch of the sound. Subsequently, luxury feeling evaluation model was obtained through multiple regression analysis. As the result, the luxury feeling of the door closing sound was evaluated to be high when the sound was softer and had lower pitch. Then, we prepared several luxury door closing sounds according to the obtained evaluation model through a filter processing and conducted subjective evaluation tests again to verify the evaluation model. The result shows that the amplitude increased sound at low frequency band under 100 Hz, which was calculated to be high luxury by the evaluation model, was actually evaluated as the best among the presented sounds through the subjective test. And the luxury sound quality evaluation method was confirmed to be useful to quantify and estimate the sound quality of the refrigerator door closing sound.

Abandoning dishonesty—A brief German comment on the state of the law after Ivey

The debate about the two-pronged Ghosh test for dishonesty has troubled academics and practitioners alike for some time. Concerns were raised about the jury’s ability to determine both the objective honesty standards and the defendant’s personal compliance with it, which might result in non-meritorious personal views allowing her to escape a dishonesty verdict. In Ivey, followed by Barton and Booth, the subjective test was abandoned altogether. The change has brought no doctrinal improvement, but instead unacceptably broadened criminal liability. Leaving the determination of a nebulous moral concept such as dishonesty to the jury is misguided, as it means determining a normative rule in the first place, which is not the jury’s role. Looking at the German law on theft and fraud as a comparator system, the paper argues that dishonesty should be abandoned and replaced by a lawfulness element to which the rules on mistake of civil law can then be applied.

Enrichment of Oesophageal Speech: Voice Conversion with Duration–Matched Synthetic Speech as Target

Pathological speech such as Oesophageal Speech (OS) is difficult to understand due to the presence of undesired artefacts and lack of normal healthy speech characteristics. Modern speech technologies and machine learning enable us to transform pathological speech to improve intelligibility and quality. We have used a neural network based voice conversion method with the aim of improving the intelligibility and reducing the listening effort (LE) of four OS speakers of varying speaking proficiency. The novelty of this method is the use of synthetic speech matched in duration with the source OS as the target, instead of parallel aligned healthy speech. We evaluated the converted samples from this system using a collection of Automatic Speech Recognition systems (ASR), an objective intelligibility metric (STOI) and a subjective test. ASR evaluation shows that the proposed system had significantly better word recognition accuracy compared to unprocessed OS, and baseline systems which used aligned healthy speech as the target. There was an improvement of at least 15% on STOI scores indicating a higher intelligibility for the proposed system compared to unprocessed OS, and a higher target similarity in the proposed system compared to baseline systems. The subjective test reveals a significant preference for the proposed system compared to unprocessed OS for all OS speakers, except one who was the least proficient OS speaker in the data set.

Objective and Subjective Evaluation of Saccadic Eye Movements in Healthy Children and Children with Neurodevelopmental Disorders: A Pilot Study

The objective of this study was to characterize saccades in children with neurodevelopmental disorders (NDDG, 17 children, age: 7–12 years) and compare them with a control group (CG, 15 children, age: 7–12 years), comparing the outcomes obtained with a subjective score system (Northeastern State University College of Optometry’s Oculomotor test, NSUCO) with the objective analysis obtained through a commercially available Eye Tracker (Tobii Eye X, Tobii, Stockholm, Sweden) and a specialized software analysis (Thomson Software Solutions, Welham Green, UK). Children from the NDDG obtained significantly lower NSUCO scores (p < 0.001) compared with CG. Concerning eye tracking analyses, we found a significantly higher number of hypometric saccades in NDGG (p ≤ 0.044). Likewise, we found a significantly higher percentage of regressions in the NDDG for a time interval of presentation of stimuli of 1 s (p = 0.012). Significant correlations were found between different NSUCO scores and percentage of regressions, number of saccades completed and number of hypometric saccades. The presence of hypometric saccades and regressions seems to be a differential characteristic sign of children with neurodevelopmental disorders that can be detected using an objective eye tracking analysis, but also using the subjective test NSUCO that can be easily implemented in all clinical settings.

Termination of Repeat Testing in Chemical Laboratories Based on Practice Guidelines: Examining the Effect of Rule-Based Repeat Testing in a Transplantation Center

Background. Although the automation of instruments has reduced the variability of results and errors of analysis, in some laboratories, repeating a test to confirm its accuracy is still performed for critical and noncritical results. However, the importance of repeat testing is not well established yet, and there are no clear criteria for repeating a test. Materials and Methods. In this cross-sectional study, all repeated tests for 26 biochemical analytes (i.e., albumin, alkaline phosphatase (ALP), alanine aminotransferase (ALT), amylase, aspartate aminotransferase (AST), bilirubin total (BT), bilirubin direct (BD), blood urea nitrogen (BUN), calcium, chloride (Cl), cholesterol total (CholT), creatine kinase (CK), creatinine (Cr), glucose, gamma-glutamyl transferase (GGT), high-density lipoprotein-cholesterol (HDL-c), iron, lactate dehydrogenase (LDH), LDL-c, lipase, magnesium (Mg), phosphorus (Ph), protein total (ProtT), total iron binding capacity (TIBC), triglyceride (TG), and uric acid) were assessed in both critical and noncritical ranges over two consecutive months (routine subjective test repeats in the first month and rule-based repeats in the second month). To determine the usefulness of test repeats, differences between the initial and verified results were compared with the allowable bias, and repeat testing was considered necessary if it exceeded the allowable bias range. All causes of repeat testing, including linearity flags, delta checks, clinically significant values, and critical values, were also documented. All data, including the cause of repeats, initial and verified results, time, and costs in the two consecutive months, were transferred to Microsoft Excel for analysis. For comparison of data between the months, Student’s t-test was used. Results. A total of 7714 repeat tests were performed over two consecutive months. Although a significant decline (38%) was found in repeated tests in the second month ( P < 0.001), there was no significant change in the percentage of unnecessary repeats (77% in the first month and 74% in the second month). In both consecutive months, AST and ALT were the most commonly repeated tests, and delta check was the most common cause of repeat testing. Mg, ALP, AST, and lipase showed the highest rates of necessary repeats, respectively (the least stable tests), while albumin, LDL, and CholT tests showed the highest rates of unnecessary repeats, respectively (the most stable tests). The total cost and delay in turnaround time (TAT) due to repeated testing decreased by 32% and 36%, respectively. Conclusion. Although repeat testing has been shown to be unnecessary in most cases, having a strict policy for repeat testing appears to be more valuable than avoiding it completely. Each laboratory is advised to establish its own protocol for repeat testing based on its own practice.

Investigating Clustering Technique on Students’ Writing Skill in Narrative Text

Abstract— The purpose of this study was to explain the effect of clustering technique toward students’ writing skill of narrative text. This study used a quasi-experimental design with post-test only control group design. The population of this study was all students of tenth graders at SMA Tulus Bhakti Bekasi, with the amount of the students were 132 students. The samples of this research were taken by cluster random sampling consisting of 60 students. The students were divided into two classes, 30 experimental class (X IPS 1) and 30 other students in the control class (X IPS 2). The instrument used to collect data was learning achievement test in form of subjective test. The statistic method used to analyze the data was One-Way ANOVA using SPSS through the testing of Null Hypothesis. The result of ANOVA showed that value of sig (significance) is lower than 0.05 (0.000<0.05) or Fobserved with df (1/58) is higher than Ftable (28.185>4.01). Accordingly, the Null Hypothesis (Ho), stating that there is no effect of clustering technique on grade ten students’ writing skill of narrative text at SMA Tulus Bhakti, was rejected and the Alternative Hypothesis (Ha), stating that there is significant effect of clustering technique on grade ten students’ writing skill of narrative text at SMA Tulus Bhakti, was accepted. Based on the analysis of the data, the research has empirically proved that there is an effect of clustering technique has an effect on grade ten students’ writing skill of narrative text at SMA Tulus Bhakti Bekasi. Keywords— clustering techniques; quasi-experimental method; writing skill