On Rural Typologies with Neural Network Method: Case Study on Xining Region

Zhang Weifeng, Zhang Li
2020 Journal of Regional and City Planning  
There are great differences between the rural areas of China, and rural areas themselves have complex development characteristics. With the implementation of the strategy of rural revitalization, 'one-fits-all' rural policy standards have had difficulty to meet the needs of different types of rural development. Rural policies adapted to local conditions cannot be separated from the identification of rural types. How to scientifically distinguish between rural areas and a widely ranged rural
more » ... logy is of great significance. This paper attempts to introduce the artificial neural network method to identify rural types and to explore the impact of the neural network model trained with different sample data on the results of rural type recognition. Taking the Xining region as an example, rural types were identified and the applicability of the model was tested. Finally, the recognition results of the neural network model were examined and further improvement of the proposed method is discussed. Abstrak. Ada perbedaan besar di daerah perdesaan di negara China, dan daerah pedesaan itu sendiri memiliki karakteristik pembangunan yang kompleks. Dengan penerapan strategi revitalisasi pedesaan, standar kebijakan pedesaan "satu-untuk-semua" sulit memenuhi kebutuhan berbagai jenis pembangunan pedesaan. Kebijakan-kebijakan perdesaan yang disesuaikan dengan kondisi lokal tidak dapat dipisahkan dari identifikasi tipe perdesaan. Cara membedakan secara ilmiah daerah perdesaan dan berbagai tipologi perdesaan menjadi sangat penting. Makalah ini memperkenalkan metode jaringan saraf tiruan untuk mengidentifikasi tipe perdesaan, dan mengeksplorasi dampak model jaringan saraf yang telah dilatih untuk mengenaili tipe perdesaan dengan data sampel yang berbeda. Dengan mengambil wilayah Xining sebagai contoh, tipe perdesaan diidentifikasi, dan penerapan model diuji. Akhirnya, hasil pengenalan oleh model jaringan saraf dievaluasi , dan peningkatan lebih lanjut dari metode ini dibahas.
doi:10.5614/jpwk.2020.31.1.2 fatcat:ja3nngx6yfeq5ktkerv5qr7p6i