This function predicts hexagonal IDs for a test set based on existing bin centroids.
Arguments
- df_bin_centroids
The training dataset containing high-dimensional data with IDs.
- nldr_df_test
The non-linear dimensionality reductions that need to find the prediction.
- x
The name of the column that contains first 2D embeddings component.
- y
The name of the column that contains second 2D embeddings component.
Examples
num_bins_x <- 4
shape_value <- 1.833091
hexbin_data_object <- extract_hexbin_mean(nldr_df = s_curve_noise_umap, num_bins_x,
shape_val = shape_value)
df_bin_centroids <- hexbin_data_object$hexdf_data
predict_hex_id(df_bin_centroids = df_bin_centroids, nldr_df_test = s_curve_noise_umap,
x = "UMAP1", y = "UMAP2")
#> # A tibble: 75 × 4
#> UMAP1 UMAP2 ID pred_hb_id
#> <dbl> <dbl> <int> <dbl>
#> 1 -2.81 -3.91 1 6
#> 2 0.959 -0.00271 2 24
#> 3 1.54 0.462 3 29
#> 4 -2.31 -5.50 4 2
#> 5 -1.76 -3.46 6 12
#> 6 1.53 5.75 7 44
#> 7 0.930 -0.175 8 24
#> 8 0.319 -1.61 9 18
#> 9 1.37 0.0541 11 24
#> 10 1.90 4.94 12 45
#> # ℹ 65 more rows