Under the background of the development of the network information age, the current Internet industry has obtained more development opportunities, but it has also brought corresponding challenges in the process of wide application. In the development and construction of modernization, society pays more attention to the supervision and determination of the characteristics of online public opinion. From the perspective of the current characteristics of network public opinion, because social information is more extensive and involves many fields, network public opinion has a high degree of complexity and diffusion. Therefore, it is necessary to strengthen the analysis and application of relevant data mining systems in order to achieve efficient management of network public opinion. The key to the disadvantage of the traditional excavation of public opinion communication characteristics lies in the lag of the excavation process, and it is difficult to deal with malignant public opinion in a timely and effective manner. Therefore, in order to truly solve the lagging problem of public opinion data dissemination feature mining technology, it is necessary to strengthen the application of artificial intelligence technology in it.

The Cisadane Watershed is in a critical state, which has expanded residential areas upstream of Cisadane. Changes in land use and cover can impact a region's hydrological characteristics. The Soil and Water Assessment Tool (SWAT) is a hydrological model that can simulate the hydrological characteristics of the watershed affected by land use. This study aims to evaluate the impact of land use change on the hydrological characteristics of the Cisadane watershed using SWAT under different land use scenarios. The models were calibrated and validated, and the results showed satisfactory agreement between observed and simulated streamflow. The main river channel is based on the results of the watershed delineation process, with the watershed boundary consisting of 85 sub-watersheds. The hydrological characteristics showed that the maximum flow rate (Q max) was 12.30 m³/s, and the minimum flow rate (Q min) was 5.50 m³/s. The study area's distribution of future land use scenarios includes business as usual (BAU), protecting paddy fields (PPF), and protecting forest areas (PFA). The BAU scenario had the worst effect on hydrological responses due to the decreasing forests and paddy fields. The PFA scenario yielded the most favourable hydrological response, achieving a notable reduction from the baseline BAU in surface flow, lateral flow, and groundwater by 2%, 7%, and 2%, respectively. This was attributed to enhanced water infiltration, alongside increases in water yield and evapotranspiration of 3% and 15%, respectively. l Therefore, it is vital to maintain green vegetation and conserve land to support sustainable water availability.