The implemented and the given dataset both clusters are obtained using rioja, It had been seen that with non-euclidian distance metric the chclust function doesn't give expected results.
The code can be found inside folder rioja_easy. The pdf stating all details can be found here: rioja
Function chaclust takes similarity matrix, ncol and width as input and a boolean(f=0) to give hclust object as output. For f = 0 it gives HeapHop function's output which has the merge data, for f= 1 it give adjClustBand_heap function's output (hclust object). The comparision is being made of rioja chclust and adjClustBand_heap functions outcomes.
The Similarity Matrix formation, the Function and comparision with rioja can be found inside folder HeapHop. The pdf stating all details can be found here: Similarity| Comparision
The function adjClustBand_heap in the package adjclust is modified from ward's criterion to single linkage criterion giving single linkage clustering of time complexity O(plogp + ph). As Distance matrix is unsorted hence we can't use divide and conquore to reduce the time complexity furthermore from O(ph).
The function descriptions and the edited files are inside the folder cpp_code. The pdf stating all details can be found here: SingleLinkage| FunctionDescription of adjClustBand_heap
In adjclust, the mainFunctions.c and adjClustBand_heap.R is updated : here