Autoclassification of products using artificial intelligence

A method comprising:

• receiving, by a processor, training data and a category hierarchy related to a product from a user device over a network, the training data including an attribute of the product;
• extracting, by the processor, a first feature and a second feature from the training data based on a plurality of statistical feature models including a first statistical model and a second statistical model, wherein the first and second features indicate the most representative features of the products among the training data for the corresponding statistical model, wherein each of the first feature and the second feature includes any one or a combination of a name of product, a brand name of the product, and a type of product, wherein the first feature is extracted using the first statistical model to and the second feature is extracted using the second statistical model, wherein the first statistical model indicates a measure of importance of attribute of the product;
• training, by the processor, a probabilistic classifier to provide a plurality of classification models including a first classification model and a second classification model, wherein the probabilistic classifier is trained using the first feature to provide the first classification model and using the second feature to provide the second classification model;
• calculating, by the processor, an accuracy score of each of the plurality of classification models, the accuracy score providing a measure of predicting a correct category belonging to a preset hierarchy of predefined categories for the product, wherein the accuracy score is calculated for each distinct category in the preset hierarchy;
• selecting, by the processor, a classification model from the plurality of classification models for each category in the preset hierarchy based on the accuracy score, the classification model being selected for classifying the product into a most relevant category of the preset hierarchy, wherein the selected classification model has a highest accuracy score for a corresponding category in the preset hierarchy;
• determining, by the processor, a probability of classifying the product into a correct category based on the selected classification model to validate a suitability of selected classification model;
• based on the determined probability, auto-classifying, by the processor, the product in a category of the category hierarchy;
• and
• providing, by the processor, results of the classification to an output device.

A non-transitory computer readable medium comprising machine executable instructions that are executable by a processor to:

• receive training data and a category hierarchy related to product from a user device over a network, the training data including an attribute of the product;
• extract a first feature and a second feature from the training data based on a plurality of statistical feature models including a first statistical model and a second statistical model, wherein the first and second features indicate the most representative features of the products among the training data for the corresponding statistical model, wherein each of the first feature and the second feature includes any one or a combination of a name of product, a brand name of the product, and a type of product, wherein the first feature is extracted using the first statistical model and the second feature is extracted using the second statistical model, wherein the first statistical model indicates a measure of importance of attribute of the product;
• train a probabilistic classifier to provide a plurality of classification models including a first classification model and a second classification model, wherein the probabilistic classifier is trained using the first feature to provide the first classification model and the second feature to provide the second classification model;
• calculate an accuracy score of each of the plurality of classification models, the accuracy score providing a measure of predicting a correct category belonging to a preset hierarchy of predefined categories for the product, wherein the accuracy score is calculated for each distinct category in the preset hierarchy;
• select a classification model from the plurality of classification models for each category in the preset hierarchy based on the accuracy score, the classification model being selected for classifying the product into a most relevant category of the preset hierarchy, wherein the selected classification model has the highest accuracy score for a corresponding category in the preset hierarchy;
• determine a probability of classifying the product into a correct category based on the selected classification model to validate a suitability of selected classification model;
• and
• based on the determined probability, auto-classify the product in a category of the category hierarchy;
• and
• provide results of the classification to an output device.