Abstract With the popularity of online learning and MOOCs, sports MOOCs are gradually receiving attention. However, there are some problems with the existing sports MOOC system, such as insufficient training effectiveness and lack of personalization in the learning process. We propose an optimization scheme based on the K-means algorithm to address these issues. This plan first collects students’ learning data, including... With the popularity of online learning and MOOCs, sports MOOCs are gradually receiving attention. However, there are some problems with the existing sports MOOC system, such as insufficient training effectiveness and lack of personalization in the learning process. We propose an optimization scheme based on the K-means algorithm to address these issues. This plan first collects students’ learning data, including learning duration, practice frequency, discussion participation, etc. Then, we use the Kmeans algorithm to divide these data into different groups and develop more personalized training plans based on the characteristics and needs of different groups. At the same time, we also considered the characteristics of the sports discipline. We incorporated various factors such as physical health, sports skills, and psychological quality into the training plan to comprehensively improve students’ sports literacy. Through experimental verification, we found that the optimization scheme based on the K-means algorithm can significantly improve the training effectiveness of the sports MOOC system. Compared with traditional training modes, the optimized mode has higher learning efficiency and lower learning costs. In addition, students also showed higher satisfaction and stronger learning motivation towards the optimized training mode. Read More Read Less

Abstract With the rapid development of information technology, document information management has become increasingly important. To improve the efficiency and accuracy of document information management, we propose a solution based on the BP neural network algorithm. This scheme first preprocesses the document information, including text cleaning, word segmentation, feature extraction, etc. Then, we used the BP neural network algorithm to classify... With the rapid development of information technology, document information management has become increasingly important. To improve the efficiency and accuracy of document information management, we propose a solution based on the BP neural network algorithm. This scheme first preprocesses the document information, including text cleaning, word segmentation, feature extraction, etc. Then, we used the BP neural network algorithm to classify and recognize the document information. Specifically, we used Multilayer Perceptron (MLP) as the model structure of the BP neural network algorithm, trained and optimized through a backpropagation algorithm. At the same time, we also used cross-validation and early stop techniques to avoid overfitting and underfitting issues. Through experimental verification, we found that the document information management system based on the BP neural network algorithm has high accuracy and efficiency. This system has higher classification accuracy and a lower false alarm rate than traditional text classification algorithms. In addition, the system also has good generalization performance and can adapt to the document information management needs of different fields. Read More Read Less

Abstract Sentiment analysis as per the textual description suggests that, it provides the sentiments or emotions on any form of given data in real life. As the social media information, now-a-days is flooded with various kinds of data from Facebook, Instagram, Twitter, and Whatsapp and so on. It is the need to handle data intelligently and classify as malicious and genuine... Sentiment analysis as per the textual description suggests that, it provides the sentiments or emotions on any form of given data in real life. As the social media information, now-a-days is flooded with various kinds of data from Facebook, Instagram, Twitter, and Whatsapp and so on. It is the need to handle data intelligently and classify as malicious and genuine data. The resultant growth in the area of social media entices tremendous challenges to the researchers in the field of social media analytics and deep learning. This paper contributes a user-friendly web application on ‘sentiment analysis of live twitter data’ using the keyword or handle, built on TextBlob library available in Python and Streamlit framework. The input data on the web application is firstly preprocessed with data cleaning, feature extraction and unstructured dataset view is filtered. Pre-processed data is further analyzed to collect the sentiments from a given twitter posts and predict into three classes viz positive, negative and neutral. The results presented in the work highlights the analysis of sentiments extracted from live tweets using keyword “Russian Ukraine World War” and classify the opinions as true or false tweets with positive, negative or neutral opinion by the use of Cat Boost Classifier. Read More Read Less