As the personal mixed ANN has several parameters that are not easy for the user to determine, including the time period of training data for the short term as well as the weighing of short-term and long-term ANN (i.e., w1 and w2), we tested, under different parameters, the proportion of the songs recommended by the personal mixed ANN using different clusters that the user could accept.
However, rather than using different clusters of organisms determined by k-means clustering, we used these pre-defined clusters of organisms.
It is worth for further investigation to try other reasonable strategies by applying prior knowledge such as pre-defined genetic pathways/networks or brain circuits, or by using different co-clustering algorithms (e.g., [26]) to take into consideration of relevant local expression patterns.
Figure 6 illustrates the result of the FCM clustering by using different colors to represent the 50 different clusters in a whole-brain DWI.
This step includes assessing the similarity of results obtained by using different sampling techniques and clustering methods.
For this purpose, SNPs were sorted consecutively by using different metrics (heterozygote excess, cluster separation, parent-parent-child errors) and those that showed extreme values were checked visually for correct clustering and, where appropriate, were manually re-clustered.
By using different network perspectives to each cluster, the addition or deletion of clusters, as well as the changes in nodes' cluster association, will only have an impact to hierarchical nearby clusters in which the changes occur.
Thereafter, expression of the most efficient gene cluster was optimized by using different promoters, including PT7, Ptac, Pc, and Pabc.
In order to show the performance of the proposed codebook design, the CDI of the selected users and clusters is quantized by using different codebooks.
First, we compute the average dissimilarities among the pixels in several clusters of different sizes, by using different window sizes.
