classification

Transportation: (1) Airworthiness of the ship is determined by a qualified inspector based on factors such as structure, materials used, dimensions, etc. The classification is given for a certain period of time and it is necessary to review again after this (or before the accident) in order to preserve the original grade. (2) assign categories to specific goods, apply appropriate category rates and relevant rules and regulations

For classification cases, we will only discuss binary classification (two classes) for simplicity. Before classifying data observations, you need to decide what is positive and what is negative. For spam or ham (ie non-spam), spam is a positive name and ham is negative. This is considered a real good factor if the model predicts that incoming mail is spam and that it is actually spam. Since the model predicts spam (positive class), the actual classes match the predictions and are therefore positive. Conversely, if the incoming mail is actually not marked as spam even though it is not spam, it is considered false positives.

Automatic classification is the process by which a training computer classifies items into one or more predefined buckets. There is no doubt that the general classification is spam filtering. Because the simplest algorithm works like a human, spam filters are relatively easy to understand. By executing this frequency analysis, various problems can be solved. A general classification task is emotional analysis - is the statement positive or negative? The company uses emotional analysis to judge whether people dislike their products on Twitter or upgrade customer service emails. Classifiers do not actually understand the abstract concept of negative nature, but comments containing words such as "broken", "disappointed", "failed" may not be tributes.

Emotion analysis or classification can be done in various ways. The amount of data you have is an important determinant of which machine learning algorithms are chosen to train emotional classification models. Recently, there has been some progress with using deep learning for classification work, but it works well enough (sorry about the large amount of data)