LSTM-CNN-CatBoost-GSSSA: Model Evaluation Results for ADSS Cable Vibration Recognition

97.6% accuracy was achieved by the LSTM-CNN-CatBoost-GSSSA model in recognizing external disturbances on urban buried optical cables. This 5.5 KB XLS file contains experimental results comparing the hybrid model against SVM, CNN, and other baselines. Author Xiaojuan Chen uploaded the evaluation data to figshare under a CC-BY-4.0 license in May 2026.

Use Cases

• Benchmarking hybrid LSTM-CNN-CatBoost models against other machine learning algorithms based on the described accuracy comparisons.
• Analyzing the performance of hyperparameter optimization using the golden sine–enhanced Sparrow Search Algorithm (GSSSA) for CatBoost.
• Studying feature fusion techniques for vibration signal recognition based on the described temporal, frequency, and spatial propagation characteristics.
• Developing real-time monitoring systems for power grid security based on the described application to ADSS optical cable damage.

Strengths

• Model achieved a reported 97.6% accuracy, surpassing six other baseline models by margins up to 15%.
• Data is shared under a permissive CC-BY-4.0 license, facilitating reuse and redistribution.
• File is a small 5.5 KB XLS, making it quick to download and inspect.

Limitations

• Row count and column definitions are unknown, requiring manual inspection to understand the data structure.
• The dataset contains only model evaluation results; the underlying vibration signal data is not included.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

figshare, uploaded by author Xiaojuan Chen.

Collection Method

Likely contains experimental results from testing the described hybrid recognition model.

Time Range

Experimental timeframe is not specified.

Freshness

Last updated 2026-05-04 17:27:10; freshness should be verified.

Geography

Focus is on urban areas, but specific locations are not provided.