System and method for generating an alternative product recommendation

#1. Field of the Invention

#The present invention is directed to systems and methods for generating product summaries. In particular, the present invention is directed to such systems and methods that allow automatic generation of product summaries together with alternative product recommendations that read naturally.

#2. Description of Related Art

#FIG. 1 is a schematic illustration of a product review generator in accordance with one example embodiment of the present invention.

#FIG. 2A is a schematic illustration showing the relationships between elements of a product summary that is generated by the product review generator of FIG. 1.

#FIG. 2B is the product summary of FIG. 2A which has been instantiated to provide a product summary that reads naturally.

#Additive rule: a rule for composing a secondary attribute, which indicates that if the primary attribute is instantiated more than once, then the values will be concatenated. See also: secondary attribute, override rule and optimizing rule. Assertion model: a set of assertion templates, which define various forms in which an assertion can manifest itself as a sentence. Assertion template: One of several grammatical patterns, together with additional field names such as the key attribute for the template, an optional excluded feature, and an optional requisite feature. The sentence pattern can be a list of “snippets.” See also: snippet, assertion, key attribute, excluded feature, requisite feature. Assertion: a point or premise of information, fact, or opinion being made by any number of possible sentences or fragments which express that point; the common state-of-affairs depicted by possible variants of an assertion. Attribute name-value pair: A matched attribute name and attribute value, also called a feature. E.g. “3-megapixel resolution.” May also include the units of the attribute name. Attribute: any defined attribute of a product or product category, whether intrinsic or extrinsic, whether primary or secondary, is considered an attribute. Connotation: Within an assertion template, an internal flag indicating whether the tone of the sentence is positive, negative, or neutral. This flag is set internally to match the connotation of the value of the key attribute of the assertion template. It is then used to determine whether a continuous transition or contrasting transition is needed when placing the sentence into a paragraph. Excluded feature: In a secondary attribute composition rule, or in an assertion template, the attribute name-value pair which, if present in the product, should cause the building of the attribute or assertion to fail. Extrinsic attribute: an attribute which is not inherently a part of the product but is dependent on external factors such as the marketplace, human judgment about the product, human purposes satisfied by the product, etc. Feature: an attribute name-value-unit combination. For example, “resolution” is an attribute name, “3 megapixel” is a value-unit pair, and “3 megapixel resolution” is a feature. Field name: any field name of an attribute, vocabulary, product, or assertion. Numerous field names can be referenced in the input files accepted by the product summary generator. Importance rating: a rating of how important an attribute is within a product category. For example, a scale of 1-10 can be used. These ratings can be modified by a usage scenario. Intrinsic attribute: an attribute not dependent on comparison with things or events outside the product itself. See also: extrinsic attribute. Key attribute: Attribute that is very important to a particular product-category and at least partially defines performance or desirability of a product in the product category. The key attribute of a category is used in special ways, such as to give context to a product in introductory remarks, for example, calling a camera a “3-megapixel camera”. Each assertion template within an assertion model has its own key attribute. Within an assertion template, any references to applicable field names are taken to refer to this key attribute. Optimizing rule: a rule for building a secondary attribute which means that if the primary attribute is instantiated more than once, then the new value will be an override if, and only if, the previously instantiated value is non-optimal; otherwise the old value can be retained. If more than one optimal value is instantiated, they can be additive with respect to each other. If at least one optimal value is instantiated, non-optimal values are over-written and no further non-optimal values will be admitted. Implies that results fields for secondary attribute composition rule contain some result value(s) marked with an appended asterisk (*) to indicate that they are optimal, while other values are not. See also, secondary attribute, override rule and additive rule. Ordinality: indication of in what order assertions should be placed within the same thematic paragraph. Assertions, which are tied in their ordinality, are randomly ordered with respect to one another. Override rule: a rule for building a secondary attribute, which means that if there are successive redundant functions for defining a secondary attribute, each instance will override the previous one. It should be rare that an editor would have redundant override rules, however. Normally, when defining multiple rules for building the same attribute, one would set this flag to either additive or optimizing. Primary attribute: An attribute that is inputted to the system, deriving either from our spec table, price feed, or product record. Distinguished from secondary attributes, which are created dynamically by the system. Primary scenario: the usage scenario, which is selected as the default assignment for a particular product. E.g. a camera best suited for professional photography. Product summary: a body of text intended to concisely state several insights or observations about a particular product. See also: theme, assertion. Property: a primary intrinsic attribute. In other words, an attribute not dependent on any contextual elements outside the product itself, therefore independent of the market, and independent of human judgment and use-case scenarios. See also: attribute, intrinsic attribute, primary attribute. Resident functions: the built-in functions for building secondary attributes. The most commonly used examples are: sum, product, quotient, absolute value, positive threshold, negative threshold, sub-string, and test. Required feature: a feature that is required for triggering of a secondary attribute rule or an assertion-to-sentence transformation. Scalarization: the ordering of values within a value range of an attribute so as to form an evaluative scale, e.g. ordering batteries from more favorable to less favorable in the order of Lithium, NiMH, NiCad, Alkaline. The scalarization is done according to an evaluative metric—meaning the question of which aspect, element, or quality of the attribute was chosen for the scalarization. For example, batteries were ordered above probably according to their charge-life and lack of “memory,” but if they were to be ordered according to ease-of-availability and/or being least-expensive, the scalarization would be different because a different evaluative metric had been adopted. Scenario: see usage scenario. Scenario glosses: phrases established for each scenario to describe the activities, settings and environments, etc. pertaining to each usage scenario. Secondary attribute: an attribute that is formed by applying a function to one or more pre-existing attributes. Distinguished from primary attributes, which are inputted from a file and derived from a function. Both primary and secondary attributes may be referenced in the rule for forming a secondary attribute. A collection of resident functions is support by the product summary generator for creating secondary attributes. Sentence: A complete, grammatical sentence formed by the generator. Every sentence is also called a variant of its assertion model. Severity value: A value expressing how seriously good or how seriously bad it is for a product to have a certain attribute value. A very low or a very high severity is indicative of an attribute of the product that is strong interest to most consumers. Snippet: a single element of an assertion template representing a sentence fragment, and including free-form text, word replacement sets, and/or field name. See also: assertion template, word replacement set, field name. Theme: The concept, which unifies multiple assertions together. See also: assertion. Units: the unit-of-measure for an attribute, e.g. megapixels. Usage scenario: An alternative set of importance ratings and scalarizations reflecting a different purpose, point-of-view, or intended usage of a type of product. Usage scenarios are inheritable according to the taxonomical structure of product categories, and they are additive so that they can be combined. Variant: a fully-formed sentence that is derived from an assertion model. After an assertion template is selected and word-replacements and value look-ups are performed, the resulting, complete sentence is a “variant” of the assertion template. See also: assertion, assertion model, assertion template. Word replacement set: The simpler way of varying phraseology on a small scale within an assertion model, whereby simple lists of alternative, fully-interchangeable words are indicated as choices. See also: assertion, assertion model, assertion template.

#FIG. 1 is a schematic illustration of a product summary generator 10 for generating narrative product summaries in accordance with one embodiment of the present invention. It will be evident from the discussion below that the preferred embodiment of the present invention provides a product summary generator which generates product summaries that sound and read naturally so that it can be easily understood by the reader and information presented therein can be used to aid in purchasing decisions. In addition, it will also be evident that the present invention may also be utilized to provide a recommendation for an alternative product.

#For clarity purposes, the present invention is also described below as applied to generating a narrative product summary with an alternative product recommendation for digital cameras. However, it should be evident that the present invention may be readily used to generate product summaries with alternative product recommendations for any products of any product category, and application to digital cameras is merely provided as one example. Moreover, various terminologies are utilized to describe the preferred embodiment of the present invention, definitions of various terms are provided in the GLOSSARY OF TERMS section set forth above to facilitate complete comprehension of the preferred embodiment.