Class expval_matrix (o2scl)¶

class expval_matrix : public o2scl::expval_base ¶

Matrix expectation value.

See expval_base for some general notes on this and related classes.

This is a similar to expval_scalar, except that it allows updating and statistics for a set of matrices en masse. The data is stored internally in ublas matrix and tensor3 objects, but the public member functions operate with template types which are compatible with any vector class which provides double &operator[]. It is assumed that each call to add() contains a new measurement for all of the matrix entries.

Report statistics

template<class mat_t, class mat2_t, class mat3_t> inline void current_avg_stats(mat_t &avg, mat2_t &std_dev, mat3_t &avg_err, size_t &m_block, size_t &m_per_block)¶

Report current average, standard deviation, and the error in the average and include block information.

Idea for Future:

This should be made const.

Avoid the copy associated with vector_slice().

template<class mat_t, class mat2_t, class mat3_t> inline void current_avg(mat_t &avg, mat2_t &std_dev, mat3_t &avg_err)¶

Report current average, standard deviation, and the error in the average.

Idea for Future:: This should be made const.

template<class mat_t, class mat2_t, class mat3_t> inline void reblock_avg_stats(size_t new_blocks, mat_t &avg, mat2_t &std_dev, mat3_t &avg_err, size_t &m_per_block) const¶: Report average, standard deviation, and the error in the average assuming a new block size.

template<class mat_t, class mat2_t, class mat3_t> inline void reblock_avg(size_t new_blocks, mat_t &avg, mat2_t &std_dev, mat3_t &avg_err) const¶: Report average, standard deviation, and the error in the average assuming a new block size.

const tensor3<double> &get_data() const¶: Return the current data for all blocks.

friend void hdf_output(o2scl_hdf::hdf_file &hf, expval_matrix &t, std::string name)¶

friend void hdf_input_n(o2scl_hdf::hdf_file &hf, expval_matrix &t, std::string &name)¶

Public Types

typedef boost::numeric::ublas::vector<double> ubvector¶

typedef boost::numeric::ublas::matrix<double> ubmatrix¶

typedef boost::numeric::ublas::vector_slice<ubvector> ubvector_slice¶

typedef boost::numeric::ublas::slice slice¶

Public Functions

expval_matrix()¶

expval_matrix(size_t rows, size_t cols, size_t n_blocks = 1, size_t n_per_block = 1)¶: Create for a vector of size n with n_blocks blocks and n_per_block points block.

virtual ~expval_matrix()¶

expval_matrix(const expval_matrix &ev)¶: Copy constructor.

expval_matrix &operator=(const expval_matrix &ev)¶: Copy constructor.

virtual void set_blocks_mat(size_t rows, size_t cols, size_t n_blocks, size_t n_per_block)¶: Set for a matrix with n_blocks blocks and n_per_block points block.

virtual void free()¶: Free allocated data (but do not change the current values of n_blocks or n_per_block)

template<class mat_t> inline void add(mat_t &val)¶: Add measurement of value val.

Protected Attributes

tensor3 vals¶: The average for each block.

ubmatrix current¶: The current rolling average.

size_t nr¶: The number of rows (zero for an empty expval_matrix object)

size_t nc¶: The number of columns (zero for an empty expval_matrix object)