# US 2014/0195500 A1 · Real-time identification of data candidates for classification based compression

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2014/0195500 A1
- Document ID: US20140195500A1
- Status: Published application
- Kind: A1
- Application number: 14/074,053
- Filed: 2013-11-07
- Published: 2014-07-10
- Earliest claimed: 2013-01-10
- Assignees: International Business Machines Corporation
- Applicants: International Business Machines Corporation
- Inventors: Jonathan AMIT; Lilia DEMIDOV; George GOLDBERG; Nir HALOWANI; Ronen I. KAT; Chaim KOIFMAN; Sergey MARENKOV; Dmitry SOTNIKOV
- Web page: https://patentgrep.com/patent/US20140195500A1

## CPC

- G06F 16/1744: {using compression, e.g. sparse files}
- G06F 16/215: Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
- G06F 16/2365: {Ensuring data consistency and integrity}
- G06F 16/285: {Clustering or classification}
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

Identification of data candidates for data processing is performed in real time by a processor device in a computing environment. Data candidates are sampled for performing a classification-based compression upon the data candidates. A heuristic is computed on a randomly selected data sample from the data candidate for determining if the data candidate may benefit from the classification-based compression. A decision is provided for approving the classification-based compression on the data candidates according to the heuristic.

## Claims

### Claim 1

A method for real-time identification of data candidates for data processing by a processor device in a computing environment, the method comprising:
sampling data candidates for performing a classification-based compression upon the data candidates, and
computing a heuristic on a randomly selected data sample from the data candidate thereby determining if the data candidates may benefit from the classification-based compression, wherein a decision is provided for approving the classification-based compression on the data candidates according to the heuristic.

### Claim 2 (depends on 1)

The method of claim 1, further including estimating if data classes are present in the data candidates.

### Claim 3 (depends on 2)

The method of claim 2, further including, if data classes are known in advance, computing the heuristic by:
for each one of the data classes, calculating an expected number of characters to be in a data class,
calculating an expected number of characters that will not belong to a predefined set of the data classes, wherein the characters that do not belong to the predefined set of the data classes are non-classifiable data, and
calculating an actual number of the characters for each of the data classes and the non-classifiable data, wherein a minimum, a maximum, an average, and a distribution of a number of consecutive characters in each of the data classes are calculated.

### Claim 4 (depends on 3)

The method of claim 3, further including,
summing a ratio between the actual number of the characters and the expected number of the characters, and then dividing the ratio by a number of the data classes that are not empty, wherein the non-classifiable data are included in the number of the data classes during the dividing, and the number of the data classes, that are not empty, have characters that belong to the class that were observed in the input, and
performing the classification-based compression on the data candidates if the ratio exceeds a threshold.

### Claim 5 (depends on 2)

The method of claim 2, further including, if data classes are not known in advance:
calculating a pair-wise histogram by counting a number of appearances for each two-consecutive characters for creating a graph representation having weighted edges, wherein a weight is the number of appearances of the two-consecutive characters,
running a clustering operation on the graph representation whereby the clustering operation filters outliers, and each cluster in the graph representation is a data class, and
executing the heuristic computed for data classes that are known in advance for performing the classification-based compression if more than one cluster is identified, wherein the classification-based compression is not performed if the clustering operation outputs a single cluster.

### Claim 6 (depends on 1)

The method of claim 1, wherein the identifying further includes:
analyzing data in a write buffer independently for specific character types, and
performing the classification-based compression if a ratio of the specific character types exceeds non-specific character types by a threshold amount.

### Claim 7 (depends on 1)

The method of claim 1, further including performing a classification identification on the data candidates during at least one single pass for both fixed data lengths and variable data lengths.

### Claim 8 (depends on 1)

The method of claim 1, wherein the data candidates are data blocks for determining intra-block classification of the data for performing the classification-based compression on the data candidates.