Interactive K-Means Clustering Method Based on User Behavior for Different Analysis Target in Medicine

Abstract

Clustering algorithm as a basis of data analysis is widely used in analysis systems. However, as for the high dimensions of the data, the clustering algorithm may overlook the business relation between these dimensions especially in the medical fields. As a result, usually the clustering result may not meet the business goals of the users. Then, in the clustering process, if it can combine the knowledge of the users, that is, the doctor's knowledge or the analysis intent, the clustering result can be more satisfied. In this paper, we propose an interactive K-means clustering method to improve the user's satisfactions towards the result. The core of this method is to get the user's feedback of the clustering result, to optimize the clustering result. Then, a particle swarm optimization algorithm is used in the method to optimize the parameters, especially the weight settings in the clustering algorithm to make it reflect the user's business preference as possible. After that, based on the parameter optimization and adjustment, the clustering result can be closer to the user's requirement. Finally, we take an example in the breast cancer, to testify our method. The experiments show the better performance of our algorithm.

Semantic categorization of Chinese eligibility criteria in clinical trials using machine learning methods

“…Text clustering is an unsupervised learning approach to partitioning unlabeled text data into meaningful groups with similar data [ 38 ], generally used for mining valuable information, such as the categories. Such task often relies on text feature representation and vector dimension reduction.…”

Knockdown of SERPINE1 reverses resistance of triple‑negative breast cancer to paclitaxel via suppression of VEGFA

“…Based on the previous studies ( 17 , 18 ), K-means clustering method was applied to calculate the different clusters of the gene interaction network.…”

Interactive Clustering

“…The basic assumption made by the authors of the papers in this category is that an interactive and collaborative process combining the strengths of both human and machine would yield better results than a process that is purely automated or purely manual. Several examples of improving the quality of the clustering results using different strategies are given in the works presented in Andrienko and Andrienko [4], Basu et al [15], Boudjeloud-Assala et al [19], Cao et al [24], Castellanos-Garzón et al [26], Choo et al [30], Dobrynin et al [38], Hadlak et al [50], Hoque and Carenini [53], Hu et al [55], Kumpf et al [64], Lai et al [66], Lee et al [67], Lei et al [68], MacInnes et al [72], Packer et al [79], Schreck et al [86], Srivastava et al [94], Turkay et al [99,101], Zhou et al [116].…”

“…We find that most of the approaches support updating the usual parameters of the clustering methods, e.g., adjust the number of clusters or the similarity threshold parameters [5,8,19,38,43,45,49,62,65,68,70,72,79,[92][93][94]101]. A somewhat unique perspective in this category is provided by Xiao and Dunham [111], where the authors note that when the whole dataset is unknown a priori, choosing an appropriate value of the number of clusters may be difficult.…”

“…Other papers incorporate the user's feedback by adjusting parameters instead of similarity functions. An example of this is presented in Lei et al [68], where weight settings in the clustering algorithm are adjusted to reflect the user's preference as much as possible. In Jiang and Canny [62], users can investigate tradeoffs among competing goals made by different parameter choices and visualize their effect on clustering results.…”