SOAPy_st.tl.TensorDecomposition

class SOAPy_st.tl.TensorDecomposition

Function:

The data of anndata type is assembled tensor and tensor decomposition is carried out

It supports the construction of high-dimensional tensors in count form using attributes in obs and the construction of high-dimensional tensors in expression value form using attributes in obs matching genes

The constructed tensor can use Z-score, 0-1 regularization or proportional form for data balancing

Tensor decomposition provides two forms of CP decomposition and tucker decomposition, both of which can choose the non-negative form of decomposition to assemble tensors of anndata type and conduct tensor decomposition

__init__()

Initialize the tensorDecomposition object.

Methods

CP_decomposition(rank[, non_negative])	Perform CP decomposition on the tensor.
__init__()	Initialize the tensorDecomposition object.
highly_variable(factor_name, top_num)
input_tensor(tensor, factor_name, ticks_list)	Set the input tensor, factor names, and supporting dictionary.
normalization(factor_name[, method])	Normalize the tensor along a specified factor.
tensor_with_gene(adata, obs_factor[, ...])	Construct a tensor using observation factors and gene expression values from an AnnData object.
tensor_with_obs(adata, obs_factor)	Construct a tensor using observation factors from an AnnData object.
tucker_decomposition(rank[, non_negative])	Perform Tucker decomposition on the tensor.

Attributes

get_tensor

Get the current tensor.