学术论文的结论（Conclusions）是作者对所从事或开展的研究项目（课题）进行的“集成”总结，它须客观地反映研究项目(课题)的价值以及其对以后研究的指导意义。

从“时空”来讲，结论（Conclusions）与引言（Introduction）遥相呼应。由于作者在引言（Introduction）部分一般要介绍研究的目的，所以在结论（Conclusions）部分要明确告知读者这些目的是否已经达到，取得了哪些结果，有何意义，还存在哪些问题，解决问题的建议，以及未来进一步研究的设想等。

学术论文结论一般包含以下内容：

1、研究过程综述。通常包括概括项目(课题)研究的内容、结果及其意义。

2、研究得出的结论。一般要比较具体阐明研究得出了什么结论，有何创造性成果或见解，有何实际应用前景等。

3、所开展的研究对该领域的启示。。

4、局限性、不足之处，特别是尚待解决的问题。

5、提出后续进一步研究的建议或设想。

例一：Landslide inventory maps document the extent of landslide phenomena in a region, and show information that can be exploited to investigate the distribution, types, pattern, recurrence and statistics of slope failures, to determine landslide susceptibility, hazard, vulnerability and risk, and to study the evolution of landscapes dominated by mass-wasting processes. Despite their importance, landslide maps remain surprisingly rare (Brabb and Harrod, 1989; Nadim et al., 2006). We argue that this is chiefly because of the difficulties and uncertainties inherent in the preparation of landslide inventories. There is a clear need for new landslide inventory maps, including geomorphological, event, seasonal and multi-temporal maps. The need exists for new, standardized landslide maps covering systematically large areas extending for several thousands of square kilometers, comprising states (Cardinali et al., 1990; Trigila et al., 2010) and even entire continents (Van Den Eeckhaut and Hervás, 2011). It is equally important to prepare inventory maps for areas where landslides are frequent and abundant, and where slope failures are sparse or rare (e.g., Van Den Eeckhaut et al., 2007, 2009). Lack of basic information on landslide distribution and abundance hampers the possibility of determining landslide susceptibility, hazard and risk at the regional, national and continental scales (e.g., Brabb et al., 2000).

The quality of the landslide inventories, which depends on the accuracy, type and certainty of the information shown in the maps, is difficult to determine, limiting the use of the inventories. New and emerging mapping methods, based chiefly on satellite, aerial and terrestrial remote sensing technologies, can greatly facilitate the production and the update of landslide maps. Review of the literature has shown that the most promising approaches exploit VHR optical, monoscopic and stereoscopic satellite images, analyzed visually or through semi-automatic procedures, and VHR digital representations of surface topography captured by LiDAR sensors. A combination of satellite, aerial and terrestrial remote sensing data represents the optimal solution for landslide detection and mapping, in different physiographic, climatic and land cover conditions. The new methods and techniques are also expected to facilitate the definition and systematic application of much needed standards for the production of landslide maps. This will have positive feedbacks on the quality of many derivative products, including hazard and risk assessments, and geomorphological investigations on the construction and dismantling of landscapes. [Fausto Guzzetti, et al. Landslide inventory maps: New tools for an old problem. Earth-Science Reviews 112 (2012) 61–62]

例二：A fully mechanized caving mining face was used as a base of in-situ test, aiming at realizing the accurate division of “three zones” and remarkable prediction of coal temperature. The conclusions obtained were as follows:

(1) Based on the monitoring data of gases and temperature in the gob of the fully mechanized caving face, the three-dimensional distribution and contour maps of the concentration of O2, CO, CO2, CH4, and temperature in the gob were acquired by griddata interpolation, which made gases and temperature in the gob more vividly and intuitively display in three-dimension. In the meantime, the variation laws of gases and temperature in the gob was systematically analyzed and explained from the perspective of coal spontaneous combustion.

(2) In order to more accurately divide the “three zones” in gob, it was proposed to comprehensively divide the “three zones” of coal spontaneous combustion using O2 concentration in the range of 5–18 vol%, the appearance and disappearance of CO, and the heating rate K = 0 °C/m. The oxidation zone was 74–119 m in the air intake side, 18–128 m in the middle of the gob, and 17–110 m in the air return side. The danger zone of coal spontaneous combustion was determined combined with the conditions of gas explosion, which was 18–56 m in the middle of the gob and 17–60 m in the air return side. The minimum mining speed was calculated to be 4.8 m/day, which ensured that spontaneous combustion did not happen in the gob.

(3) PSO-SVR model was constructed to predict the temperature of coal spontaneous combustion based on the information of gases concentration in the gob and distance from the measuring points to the working face. A comparison between PSO-SVR and standard SVR, BPNN, and MLR model revealed that the nonlinear fitting capability of the MLR model was the worst, and the BPNN model was prone to have the phenomenon of over-fitting. It outperformed other models with respect to the training samples, but its capacity of generalization and popularization was inferior to other models. Compared with standard SVR model, PSO-SVR model had better capacity of generalization and robustness, and the prediction accuracy was higher, which indicated that the intelligent algorithm had the potential to improve the prediction precision.[Jun Deng, et al. Determination and prediction on “three zones” of coal spontaneous combustion in a gob of fully mechanized caving face. Fuel 211 (2018) 467–468]