Social media multi-modal rumor detection faces challenges such as weak cross-modal feature correlation and insufficient intrinsic representation of data. Therefore， a rumor detection method based on cross-modal attention mechanism and contrastive learning was proposed. In the method， fine-grained features of text and vision were extracted by a multi-modal feature module， cross-modal co-attention mechanism and discriminative learning were utilized to enhance inter-modal correlation， complex semantic contexts were captured by using multi-head self-attention， and a contrastive learning module was introduced innovatively to achieve feature optimization under machine supervision. Experimental results on the public Twitter-16 and Weibo datasets show that the accuracy of the proposed method is improved by 5.47 and 4.44 percentage points， respectively， compared with that of the existing optimal model MMFN （Multi-Modal Fusion Network）， verifying the key roles of fine-grained feature mining and cross-modal similarity modeling in detection performance. It can be seen that analyzing multi-modal content differences deeply and strengthening cross-modal association mechanism can improve the recognition accuracy of social media rumors effectively.

Due to the disadvantages such as sensitive to the initial selection of the center, slow convergent speed and poor accuracy in traditional K-medoids clustering algorithm, a novel K-medoids algorithm based on improved Granular Computing (GrC), granule iterative search strategy and a new fitness function was proposed in this paper. The algorithm selected K granules using the granular computing thinking in the high-density area which were far apart, selected its center point as the K initial cluster centers, and updated K center points in candidate granules to reduce the number of iterations. What's more, a new fitness function was presented based on between-class distance and within-class distance to improve clustering accuracy. Tested on a number of standard data sets in UCI, the experimental results show that this new algorithm reuduces the number of iterations effectively and improves the accuracy of clustering.

Nutch crawling performance was optimized by tunning Nutch MapReduce job configurations. In order to optimize Nutch performance, firstly Nutch crawling processes were studied from the view of Hadoop. And based on that, the characters of Nutch jobs workflows were analyzed in detail. Then tunned job configurations were generated by profiling Nutch crawling process. The tunned configurations were set before the next job running of the same type. The appropriate profiling interval was selected to consider the balance between cluster environmental error and profiling load, which improved optimization result. The experimental results show that it is indeed more efficient than the original programs by 5% to 14%. The interval value of 5 makes the best optimization result.