This paper provides a comprehensive survey of deep learning methods for named entity recognition (NER). NER aims to identify and classify named entities in unstructured text, which is crucial for knowledge graph construction. Traditional NER methods rely heavily on manual feature engineering and struggle with adapting to large datasets in complex linguistic contexts. Recent advances in deep learning have led to the development of numerous NER methods. The paper first introduces the problem definition and limitations of traditional methods. It then presents commonly used NER datasets and categorizes them based on the complexity of named entities. The paper summarizes typical deep learning-based NER methods according to the development history of deep learning models. It conducts an in-depth analysis and comparison of methods achieving outstanding performance on representative datasets. The paper also reproduces and analyzes the recognition results of some typical models on three different types of datasets. Finally, it offers insights into the future trends of NER development. The paper addresses the limitations of previous NER reviews, such as being oriented to a single background and lacking visualization of data and experiments. It summarizes deep learning-based NER research in five aspects: concepts, data, methods, experiments, and future work. The main contributions include providing researchers with a visual understanding of the NER process, classifying NER data into three types based on entity complexity, and summarizing methods according to the development history of deep learning. The paper also presents the results of recent NER works for three different classes of NER data, enabling the performance trends of deep learning-based NER methods to be quantitatively demonstrated. It experiments with different data using methods oriented to three types of NER research and compares and analyzes the visualization results of these methods for different types of NER. Finally, it summarizes the challenges and future research directions of deep learning-based NER methods, providing further research motivation for NER research. The paper is structured as follows: Section 2 provides a brief history of NER. Section 3 introduces some common datasets and the main research difficulties for NER. Section 4 highlights and incorporates the more prevalent deep learning-based NER methods used in recent years. Section 5 briefly discusses potential future study focuses.This paper provides a comprehensive survey of deep learning methods for named entity recognition (NER). NER aims to identify and classify named entities in unstructured text, which is crucial for knowledge graph construction. Traditional NER methods rely heavily on manual feature engineering and struggle with adapting to large datasets in complex linguistic contexts. Recent advances in deep learning have led to the development of numerous NER methods. The paper first introduces the problem definition and limitations of traditional methods. It then presents commonly used NER datasets and categorizes them based on the complexity of named entities. The paper summarizes typical deep learning-based NER methods according to the development history of deep learning models. It conducts an in-depth analysis and comparison of methods achieving outstanding performance on representative datasets. The paper also reproduces and analyzes the recognition results of some typical models on three different types of datasets. Finally, it offers insights into the future trends of NER development. The paper addresses the limitations of previous NER reviews, such as being oriented to a single background and lacking visualization of data and experiments. It summarizes deep learning-based NER research in five aspects: concepts, data, methods, experiments, and future work. The main contributions include providing researchers with a visual understanding of the NER process, classifying NER data into three types based on entity complexity, and summarizing methods according to the development history of deep learning. The paper also presents the results of recent NER works for three different classes of NER data, enabling the performance trends of deep learning-based NER methods to be quantitatively demonstrated. It experiments with different data using methods oriented to three types of NER research and compares and analyzes the visualization results of these methods for different types of NER. Finally, it summarizes the challenges and future research directions of deep learning-based NER methods, providing further research motivation for NER research. The paper is structured as follows: Section 2 provides a brief history of NER. Section 3 introduces some common datasets and the main research difficulties for NER. Section 4 highlights and incorporates the more prevalent deep learning-based NER methods used in recent years. Section 5 briefly discusses potential future study focuses.