Public opinion can spread like wildfire on campus social networks. Traditional methods of monitoring and searching these networks can be slow and wasteful. To solve this problem, researchers have introduced a new approach. They've combined dynamic deletion with the shortest path algorithm to classify network texts more efficiently. This method helps to prevent resource waste.

Next, they've built a unique text sentiment classification model using a mix of convolutional neural networks and recurrent neural networks. But wait, there's more! They've also added an attention mechanism to improve the model's classification power.

The results? The dynamic deletion-shortest path algorithm had impressive scores: an average precision rate of 97. 30%, a recall rate of 79. 55%, and an F-value of 87. 53%. Plus, it worked at a speed of 397 KB/s, outperforming traditional methods.

When it came to classifying long texts, the variant recurrent neural network model showed high accuracy and F-values, both over 84%. The addition of the attention mechanism boosted the text sentiment classification model's accuracy by 3. 89%.

In simple terms, the new dynamic deletion-shortest path algorithm and the sentiment classification model with the attention mechanism are game-changers. They offer superior performance and can be incredibly useful for making decisions about opinion risks in campus social networks.

Improving Campus Social Networks: A New Way to Track Opinions

Actions