validateSpectraParams

C++ Function Reference

1 Signature

std::tuple< int, int > BigDataStatMeth::validateSpectraParams(int n, int k, int ncv)

2 Description

Validate and adjust Spectra parameters for convergence.

3 Parameters

n (int): Matrix size
k (int): Number of eigenvalues requested
ncv (int): Number of Arnoldi vectors

4 Returns

Adjusted parameters that satisfy Spectra constraints

5 Details

nMatrix size kNumber of eigenvalues requested ncvNumber of Arnoldi vectors Adjusted parameters that satisfy Spectra constraints

6 Caller Graph

7 Source Code

Implementation

File: inst/include/hdf5Algebra/matrixEigenDecomposition.hpp • Lines 72-92

inline std::tuple<int, int> validateSpectraParams(int n, int k, int ncv) {
        // Ensure k is reasonable
        if (k <= 0) k = std::min(n, 6);
        if (k >= n) k = n - 1;
        
        // Calculate optimal ncv following RSpectra defaults
        if (ncv <= 0) {
            ncv = std::min(n, std::max(2 * k + 1, k + 2));
        }
        
        // Enforce Spectra constraints: k + 2 <= ncv <= n
        ncv = std::max(ncv, k + 2);
        ncv = std::min(ncv, n);
        
        // Additional safety: ensure we have enough space for convergence
        if (ncv - k < 2) {
            ncv = std::min(n, k + 2);
        }
        
        return std::make_tuple(k, ncv);
    }

8 Usage Example

#include "BigDataStatMeth.hpp"

// Example usage
auto result = validateSpectraParams(...);

---
title: "validateSpectraParams"
subtitle: "C++ Function Reference"
---

## Signature

```cpp
std::tuple< int, int > BigDataStatMeth::validateSpectraParams(int n, int k, int ncv)
```

## Description

Validate and adjust Spectra parameters for convergence.

## Parameters

- `n` (`int`): Matrix size
- `k` (`int`): Number of eigenvalues requested
- `ncv` (`int`): Number of Arnoldi vectors

## Returns

Adjusted parameters that satisfy Spectra constraints

## Details

nMatrix size kNumber of eigenvalues requested 
 ncvNumber of Arnoldi vectors Adjusted parameters that satisfy Spectra constraints

## Caller Graph

::: {.diagram-section}
<object type="image/svg+xml" data="images/namespace_big_data_stat_meth_afb53700b444c5257b41f07f12a84bdaa_icgraph.svg" class="class-diagram">
  <img src="images/namespace_big_data_stat_meth_afb53700b444c5257b41f07f12a84bdaa_icgraph.svg" alt="Function dependencies"/>
</object>
:::

## Source Code

::: {.callout-note collapse="true"}
### Implementation

**File:** `inst/include/hdf5Algebra/matrixEigenDecomposition.hpp` • **Lines 72-92**

```cpp
inline std::tuple<int, int> validateSpectraParams(int n, int k, int ncv) {
        // Ensure k is reasonable
        if (k <= 0) k = std::min(n, 6);
        if (k >= n) k = n - 1;
        
        // Calculate optimal ncv following RSpectra defaults
        if (ncv <= 0) {
            ncv = std::min(n, std::max(2 * k + 1, k + 2));
        }
        
        // Enforce Spectra constraints: k + 2 <= ncv <= n
        ncv = std::max(ncv, k + 2);
        ncv = std::min(ncv, n);
        
        // Additional safety: ensure we have enough space for convergence
        if (ncv - k < 2) {
            ncv = std::min(n, k + 2);
        }
        
        return std::make_tuple(k, ncv);
    }
```
:::

## Usage Example

```cpp
#include "BigDataStatMeth.hpp"

// Example usage
auto result = validateSpectraParams(...);
```