(4) During the past decade, several algorithms have been developed for both the normalization and summarization of raw data (robust multichip algorithm, MAS5, PLIER, etc) and to obtain differentially expressed gene lists (fold-change thresholds, multiple-testing corrected t-test, rank product, significance analysis of microarrays, linear models for microarray data, etc).
It was developed for evaluation of dynamic video summaries, tailored to category-specific summarization, and it includes annotations of important and semantically consistent segments.
We also developed summarization methods for aggregating distributed associated text, removing redundant text, and automatically generating a structured text summary for every figure [13], [14].
We select MEAD as it has been used extensively both as a platform for developing summarization systems and also as a baseline system for testing novel summarizers.
In the recent years researchers have given much attention towards developing document summarization techniques.
We have developed methods for systematic cell identification and summarization of relatively homogeneous cultures (MIN6 cells) and heterogeneous cultures (human pancreatic endocrine tissue), as well as statistical measures for mitigating plate-based artifacts on the effects of natural extracts on cell proliferation, apoptosis and expression of insulin and pdx1 in these cells.
The centerpiece of this work is an objective function for summarization that I call "maximum coverage".
Fig. 8 Multiple representations for summarization.
Single-document summarization is easy to handle since only one text document needs to be analyzed for summarization, whereas handling multi-document summarization is a complex and difficult task.
The impact of clustering for summarization of large text collection is also demonstrated in this work.
ROUGE unigram gives better results for summarization as compared to the ROUGE bigram approach.
