Is the bias(weight) obtained for each bin by the kr method of hicCorrectMatrix a division bias(weight)?

[oganesson76]

I hear that the bias(weight) obtained using the KR method is a divisive bias(weight), whereas the bias(weight) obtained using the ICE method is a multiplicative bias(weight). For instance, if the bias(weight) for bin1 is 0.1 and the bias(weight) for bin2 is 0.3, and the raw contact count for bin1 and bin2 is 100, then under the ICE method, the corrected count should be 100 × 0.1 × 0.3. Under the KR method, the corrected count should be 100 / (0.1 × 0.3).I want to know whether the kr method and ice method of hicCorrectMatrix also satisfy the above conditions.