Classification of movie reviews using term frequency-inverse document frequency and optimized machine learning algorithms

American literature news narration based on computer web technology

Driven by internet technology, online has become the main way of news dissemination, but redundant information such as navigation bars and advertisements affects people's access to news content. The research aims to enable users to obtain pure news content from redundant web information. Firstly, based on the narrative characteristics of literary news, the Term Frequency-Inverse Document Frequency (TF-IDF) algorithm is employed to extract pure news content from the analyzed web pages. The algorithm uses keyword matching, text analysis, and semantic processing to determine news content's boundaries and key information. Secondly, the news text classification algorithm (support vector machine, K-nearest neighbor, AdaBoost algorithm) is selected through comparative experiments. The news extraction system based on keyword feature and extended Document Object Model (DOM) tree is constructed. DOM technology analyzes web page structure and extracts key elements and information. Finally, the research can get their narrative characteristics by studying the narrative sequence and structure of 15 American literary news reports. The results reveal that the most used narrative sequence in American literary news is sequence and flashback. The narrative duration is dominated by the victory rate and outline, supplemented by scenes and pauses. In addition, 53.3% of the narrative structures used in literary news are time-connected. This narrative structure can help reporters have a clear conceptual structure when writing, help readers quickly grasp and understand the context of the event and the life course of the protagonists in the report, and increase the report's readability. This research on the narrative characteristics of American literature news can provide media practitioners with a reference on news narrative techniques and strategies.

Comparisons of deep learning and machine learning while using text mining methods to identify suicide attempts of patients with mood disorders

BACKGROUND

Suicide attempt is one of the most severe consequences for patients with mood disorders. This study aimed to perform deep learning and machine learning while using text mining to identify patients with suicide attempts and to compare their effectiveness.

METHODS

A total of 13,100 patients with mood disorders were selected. Two traditional text mining methods, logistic regression and Support vector machine (SVM), and one deep learning model (Convolutional neural network, CNN) were adopted to perform overall analysis and gender-specific subgroup analysis of patients to identify suicide attempts. The classification effectiveness of these models was evaluated by accuracy, F1-value, precision, recall, and the area under Receiver operator characteristic curve (ROC).

RESULTS

CNN's results were greater than the other two for all indicators except recall which was slightly smaller than SVM in male subgroup analysis. The accuracy values of the CNN were 98.4 %, 98.2 %, and 98.5 % in the overall analysis and the subgroup analysis for males and females, respectively. The results of McNemar's test showed that CNN and SVM models' predictions were statistically different from the logistic regression model's predictions in the overall analysis and the subgroup analysis for females (P < 0.050).

LIMITATIONS

A fixed number of features were selected based on document frequency to train models; this was a single-site study.

CONCLUSIONS

CNN model was a better way to detect suicide attempts in patients with mood disorders prior to hospital admission, saving time and resources in recognizing high-risk patients and preventing suicide.

A semantic similarity analysis of multiple English translations of The Analects: Based on a natural language processing algorithm

Working from the readers’ perspective, this study first investigates the online acceptance of the complete English translations of The Analects by investigating the number of online comments, downloads, academic citations, and other factors, and it ranks the different English versions according to how well they are received. The complete English translations of The Analects by D. C. Lau, James Legge, and 15 other translators are found to be well received by readers on mainstream online platforms. Then, based on five natural language processing (NLP) algorithms (TF-IDF, Word2Vec, GloVe, BERT, and SimHash), the 15 well-received English versions of The Analects are taken as samples to calculate semantic similarity. By comparing the semantic differences among the texts, this study analyzes the factors that affect the diversification of translated texts. (1) The influence of Chinese annotation on the translation semantics is great, even the greatest among many influential factors; and (2) different translators’ identities, the translation era, the translation purpose, and the translation background do not significantly affect the semantic influence of the translation. On the one hand, the readers can understand the differences between the different translations and choose an appropriate translation for their reading and learning more effectively. On the other hand, using the algorithms of NLP, we focus on the semantic similarity of different English translations of The Analects and analyze them to show the semantic differences quantitatively, which makes the comparison more intuitive and efficiently. Such a quantitative presentation of the results draws scholars’ attention to the differences in the translations.