Angular Histograms: Frequency-Based Visualizations for Large, High Dimensional Data

Question

1. What have the authors contributed in "Angular histograms: frequency-based visualizations for large, high dimensional data" ?

2. What have the authors stated for future works in "Angular histograms: frequency-based visualizations for large, high dimensional data" ?

3. How can the data density be conveyed?

4. How many bins are needed for the vector frequency?

Accepted Answer

In this paper, the authors propose two novel solutions, namely, angular histograms and attribute curves.. The authors demonstrate the results on a wide variety of data sets including real-world, high-dimensional biological data.

Accepted Answer

In the future, the authors plan to run a comparative user study for the evaluation of their techniques and further comparison with other techniques.

Accepted Answer

The data density can be conveyed by the luminance, where high density is mapped to more luminance and low density is mapped to less luminance.

Accepted Answer

For a data set containing n dimensions and m records, with each of its attributes uniformly divided into k intervals, the authors need to construct (n− 1))k2 bins for storing the line frequency [20], whereas only 4nk bins for the vector frequency.

Accepted Answer

In order to enhance the clustering and outlier effect, the line-base histogram [20] utilizes different filters, such as Gaussian or Median filters.

Accepted Answer

The binned parallel coordinates are used for context views and the traditional polyline-based parallel coordinates are used for focus views.

Accepted Answer

the clutter reduction methods on large data sets can be categorized as: alphablending, clustering, focus+context and frequency and density plots.

Accepted Answer

In order to visualize the vector aggregations, at least two features, namely direction and magnitude, have to be encoded at the same time.

Accepted Answer

From the perspective of visualization, the histogram is a visual abstraction that aggregates the univariate data, where the height of the histogram bar is mapped to only one variable or feature.

Accepted Answer

The angular brushing [11] is effective in revealing the relations between the neighboring data attributes by filtering the slopes of the line segments.

Accepted Answer

In this respect, their vector-based histogram has the advantage of taking up much less processing time and memory space than the line-basedhistogram, especially for data sets with high dimensions.

Accepted Answer

Considering the angular histogram provides a visual aggregation of the vector directions, it can be filtered by the angles of the histogram bars.

Accepted Answer

By introducing the angular information from the polyline-axis intersections, their angular histograms and attribute curves are able to depict the relationship across the data attributes.

Accepted Answer

The logarithmic histogram falls short in preserving the very low density histogram bins and the attribute curves could complement this and indicate the change rate of the data plots between adjacent axes.

Accepted Answer

If the authors calculate the average angle of these vectors, their positive and negative slopes cancel each other and lead to a flat histogram angle.

Accepted Answer

Similar to some other information visualizations [5, 6, 17, 24, 28], parallel coordinates suffers from visual clutter caused by large data sets rendered on limited screen space.

Accepted Answer

These values are not preserved in the logarithmic angular histograms shown in the middle of Figure 9, because some low-frequency histogram bins suffer from the low resolution of the display space and visually they just look the same as empty bins.

Angular Histograms: Frequency-Based Visualizations for Large, High Dimensional Data

Chat with Paper

AI Agents for this Paper

Most frequently asked questions

1. What have the authors contributed in "Angular histograms: frequency-based visualizations for large, high dimensional data" ?

2. What have the authors stated for future works in "Angular histograms: frequency-based visualizations for large, high dimensional data" ?

3. How can the data density be conveyed?

4. How many bins are needed for the vector frequency?

5. What filters are used to enhance the clustering and outlier effect?

6. What are the common uses of binned parallel coordinates?

7. What are the clutter reduction methods on large data sets?

8. How many features have to be encoded in order to visualize the vector aggregations?

9. What is the definition of the histogram?

10. What is the effect of angular brushing on the neighboring data attributes?

11. What is the advantage of the vector-based histogram?

12. What is the purpose of the angular histogram?

13. How do the authors use the angular histograms and attribute curves?

14. What is the way to show the change rate of the plots?

15. What is the average angle of the two histograms?

16. What is the drawback of parallel coordinates?

17. Why are the values not preserved in the logarithmic angular histograms?

Figures

Citations

Data Mining - Concepts and Techniques.

A survey on information visualization: recent advances and challenges

Visualizing High-Dimensional Data: Advances in the Past Decade

State of the Art of Parallel Coordinates

Visual Analytics Towards Big Data

References

Data Mining: Concepts and Techniques

Data Mining - Concepts and Techniques.

The eyes have it: a task by data type taxonomy for information visualizations

Data Mining: Concepts and Techniques

Parallel coordinates: a tool for visualizing multi-dimensional geometry

Related Papers (5)

State of the Art of Parallel Coordinates

Parallel coordinates: a tool for visualizing multi-dimensional geometry

The plane with parallel coordinates

Hyperdimensional Data Analysis Using Parallel Coordinates

Visual clustering in parallel coordinates