Traditional Culture Encyclopedia - Traditional stories - The important influence of big data on thinking mode

The important influence of big data on thinking mode

1, full sample instead of sampling: in the past, due to the limitation of data storage and processing capacity, scientific classification usually adopted sampling, that is, a part of sample data was extracted from the complete set of data, and the overall characteristics of the complete set of data were inferred through the analysis of the sample data. Now, the arrival of the era of big data provides us with the storage and processing of massive data. Therefore, with the support of big data technology, scientific analysis can be completely analyzed in a complete data set, and the results can be obtained quickly.

2. Efficiency rather than accuracy: In the past, the sampling analysis method was used in scientific analysis, which required higher accuracy, because the sampling method only aimed at the analysis of some samples, and the analysis results would amplify the error after being applied to the complete set of data. In other words, many small mistakes will cause big mistakes. This leads the traditional algorithm to pay more attention to improving accuracy rather than efficiency. In the era of big data, the full sample analysis method is adopted instead of sampling method, so there is no error amplification. Therefore, big data has the characteristic of "second-level correspondence", which requires massive data to be analyzed in a few seconds to get real-time results, otherwise it will lose data value.

3. Related rather than causal past. The purpose of data analysis, on the one hand, is to explain the development mechanism behind things; On the other hand, it is used to predict what may happen in the future. Now in the era of big data, causality is no longer important, and people turn to "correlation" instead of causality.