Patent searching and analysis have proven their utility in taking strategic business decisions, new acquisitions, exploring new geographies, and others. With patent data increasing at an unprecedented rate, precision patent searching and analysis has become a necessity. There are no set methods for patent searching and insight mining, but here are some of the most common ways to approach them:
Classification based patent search
Classification-based patent search is the most common and efficient patent search method. Classifications are standardized and globally acknowledged across global patent offices. Classification-based search comes across as one of the most reliable and widely acknowledged methods. This search strategy is particularly effective in patent spaces from various countries where precise translations are necessary, which keyword-based searches typically do not provide.
Predominantly, there are four classification models:
- The US Classification System (USPC)
- European Classification System (ECLA)
- International Patent Classification System (IPC)
- Japan’s F-Term
ECLA is a former patent classification system by the European Patent Office (EPO). ECLA has now been replaced by the Cooperative Patent Classification (CPC) system, which has been jointly developed by the EPO and USPTO. CPC, primarily used in the US and Europe, is a combination of ECLA and USPTO models and is also a superset of IPC classifications. Each of these models has its own unique advantages and provides organized ways to conduct a patent search.
Keyword-based searching has been popularized across geographies because, by far, it is the easiest way of searching. However, with the level of complexity involved, the inclusion of synonyms, semantics, and inter-operable search terms has become standard practice.
Clustering patents while searching
Patent clustering can impact patent searching and analysis in a positive way. The clustering of patents can be done either on the basis of classification codes or concepts or technology. Additional filters like Assignee and geography-based clustering also help in focusing the search query on specifics.
Clustering, however, is a complex process as it involves the categorization of data into multiple overlapping clusters and manual clustering is often tedious and time-consuming. Recently, NLP-equipped AI-based tools have been assisting searchers and analysts in segregating the patent data and clustering it for further analysis.
XLSCOUT, one of the emerging tools in the domain, offers modules equipped with Natural Language Processing (NLP) technologies that enable seamless clustering of patents by creating automated and human-assisted (Hybrid) taxonomies.
Categorization through tool-generated (automated) taxonomy
This taxonomy or clustering is done through advanced NLP concepts tagged and trained on a large patent data set. The training set provides a window for tools to create citations that act as reference points for further classification.
Machine assisted categorization on manual taxonomy
This allows users to create manual categories and then clustering the data set based on the user-created taxonomies. The tool applies NLP technology to analyze patents, identify concepts, and then cluster them according to the provided categories.
Using the right Boolean expressions
Boolean operators provide an edge to further optimize the search query and direct it to the specifics users are interested in. Along with the common Boolean operators like AND, OR, and NOT, operators like NEAR, ADJ, and BTWN are commonly used by expert searchers.
When Boolean operators are used in sync with classification-based search strings and keyword-based search strings, they often yield the desired results.
Add an extra layer of search
Automation undoubtedly helps in giving the search process more cushion, but it goes beyond that. These tools can facilitate Prior Art search because of their higher precision, the ability to segment the problem statement into smaller buckets, and considerably lesser turnaround time.
Even with machine advancement and technological supremacy, artificial intelligence or machine learning is not a replacement for human intelligence. The technology only aids in analyzing the bulk data and breaking it down into smaller chunks for further analysis.
XLSCOUT supports modules and tools that allow users to get a kick start and conduct a first pass AI-assisted analysis with its Novelty Checker tool. Novelty Checker, an AI-backed tool for first-pass patentability search, enables users to search for prior art by inputting their idea in natural language. Integrated with Para-BERT technology, Novelty Checker understands the context of each paragraph in a patent document to provide relevant results. The Novelty Checker incorporates Reinforcement Learning for users to quickly generate an automated novelty and patentability search report by selecting the top 10 or 20 results. The Novelty Checker prior art search reports include a list of results along with relevant text mapped with the key features of the invention for quick decision-making.
Using Trend analysis and data visualization:
Even after analysis, it takes a lot of effort to interpret and present textual data. Henceforth, graphical interpretation, data visualization, and tabular analysis come into play.
There are intelligent tools that can do graphical patent analysis on large patent datasets and derive insights with immediate commercial relevance. Customers can modify interactive charts from XLSCOUT to their specific needs for graphical patent analysis.
XLSCOUT offers multiple AI-assisted solutions that enable your patent searches and other IP workflows to be more efficient. These are simple to integrate with current systems, which will further streamline the procedures.