geometrical_measures: GAPGOM internal - Geometrical Similarity Measures

In Berghopper/GAPGOM: GAPGOM (novel Gene Annotation Prediction and other GO Metrics)

Description

These functions are internal functions and should not be called by the user.

Usage

sobolev_metric(x, y)

fisher_metric(x, y)

Arguments

x	gene expression value vector (numeric) of reference/novel gene that needs to be compared to the dataset.
y	also a gene expression value vector, the idea here is that indexes between x and y correspond for same or similar tissues for accurate similarity. x and y may also be swapped.

Details

These geometric functions will calculate the distance between two random sets of variables. In our usecase this would be an expression values (FPKM). Given two gene expression value lists, semantic similarity will be calculated between them. This is then returned. Fisher metric is based on the G. Lebanon et al. implementation. [1] The sobolev metrix is based on the T. Villmann et al. implementation. [2]

Value

score of similarity between the two vectors (type=double)

Notes

Both sobolev_metric() and fisher_metric() should have exactly the same input and output types. Functions are used in predict_sobolev() and predict_fisher().

References

[1]. Villmann T: Sobolev metrixs for learning of functional data - mathematical and theoretical aspects. In: Machine Learning Reports. Edited by Villmann T, Schleif F-m, vol. 1. Leipzig, Germany: Medical Department, University of Leipzig; 2007: 1-13.

[2]. Lebanon G: Learning riemannian metrics. In: Proceedings of the Nineteenth conference on Uncertainty in Artificial Intelligence; Acapulco Mexico. Morgan Kaufmann Publishers Inc. 2003: 362-369

Berghopper/GAPGOM documentation built on July 2, 2020, 11:57 p.m.