User:Jackjackkimkim/sandbox

Computational History -- Jack, James, Yamato

Computational History (not to be confused with computation history), sometimes also called Histoinformatics, is a multidisciplinary field that studies history through machine learning and other data-driven, computational approaches. (UP TO HERE IS THE ORIGINAL PAGE)

(NEW CONTENT)

One such example involves utilizing text mining to analyze how history has been recorded in the past. Advancements in computational techniques have opened up new fields such as computational linguistics and computational social science while minimizing the necessary manpower and allowing historians to parse through massive amounts of data. As computational history involves utilizing emerging computational techniques to view history, creative and new approaches are constantly created. A study by a team at Hokkaido University created a new way of distinguishing the historical importance of a figure by utilizing an algorithm similar to PageRank that examined Wikipedia links found on the page of a historical figure. Other computational techniques that are related to the field of computational history include methods from topic detection and tracking (TDT) to temporal analysis and new techniques are steadily being incorporated into the field of history. Temporal analysis with machine learning algorithms have been able to organize documents into a hierarchical system based on time by using a technique called temporal tagging. Temporal tagging analyzes a document to establish a timeline for an event by detecting relative, implicit, and explicit temporal descriptions.

Another use of computational history is that it can be used for checking the credibility of a historical source since historical narratives are often crafted to fit the political and identity narratives of those in power. Harnessing the advantages of machine learning algorithms, digging through large sets of historical documents can assist in a thorough analysis of historical evidence by finding other pieces of documents that can corroborate the claims of the original source.

Text Mining

Text mining is the process of deriving information from text. It involves the extraction of information from sources such as books, articles, and websites. The earliest use of text mining in the field of humanities is often credited to Robert Busa, who conceived of and oversaw the Index Thomisticus project. Busa’s project digitized 179 texts of Thomas Aquinas over 34 years, which totaled over 10 million words.

With the proliferation of computers and computational techniques over the last few decades, humanities researchers have increasingly been using computational methods to assist them with textual analysis. In addition to simple methods like lexical analysis, modern humanities researchers have increasingly used a combination of statistical and machine learning techniques to carry out textual analyses. Scholars commonly employ methods such as text categorization, text clustering, and sentiment analysis.

Natural Language Processing of Historical Texts

While many types of text mining, i.e. biomedical, are motivated by commercial applications, Natural Language Processing of historical texts is often motivated by humanities-based research. The first step for computational analysis of historical documents is to digitize the journals and books held at museums and libraries and there are several partnerships engaging in such work. For example in 2004, Stanford, the University of Michigan, Harvard, Oxford, and the New York Public Library partnered with the Google Library Project in a digitization project.

After text has been translated into a digital format, it must then be translated into a format that is capable of being processed by a computer. One primary approach is to map characters of a script to a set of positive integers, which gives you an encoded character set, the standard encoded set being ASCII.

Modeling the Past

There are large amounts of untapped information in historical texts, and one way to make use of them is by a technique known as "predicting the past". Researchers can better evaluate current models by overlaying current analytics with historical information derived from natural language processing of historical documents.

While drawing retrodictions or predictions about unexplored parts of history is a natural consequence of exploring the past, this process remains controversial. Using statistics to explain human behavior is said to not fully provide an explanation or a level of understanding often considered necessary in research disciplines.

Temporal Information Retrieval

Recent scholars have begun to use text analysis methods to exploit the temporal information present in historical documents. Such scholars use natural language processing to extract temporal expressions from documents, on which they perform exploratory searches, clustering, similarity studies, and other analyses. Temporal techniques give scholars the ability to identify temporal-based trends in the documents they study.

Machine Learning Applications

In recent years, machine learning methods have gained greater prevalence in the field of history. One area where machine learning has been applied is in artifact reconstruction. By training models to learn the process of degradation, for example, scholars have used AI models to reassemble eroded artifacts and damaged documents.

Edits - Claudia Ng
Computational History -- Jack, James, Yamato

Computational History (not to be confused with computation history), sometimes also called Histoinformatics, is a multidisciplinary field that studies history through machine learning and other data-driven, computational approaches. (UP TO HERE IS THE ORIGINAL PAGE)

(NEW CONTENT)

One such example involves utilizing text mining to analyze how history has been recorded in the past. Advancements in computational techniques have allowed historians to parse through massive amounts of data while minimizing the manpower required and have opened up new fields of studies like computational linguistics and computational social science. As computational history involves utilizing emerging computational techniques to view history, creative and new approaches are constantly created. A study by a team at Hokkaido University created a new way of distinguishing the historical importance of a figure by utilizing an algorithm similar to PageRank that examined Wikipedia links found on the page of a historical figure. Other computational techniques that are related to the field of computational history include anywhere from topic detection and tracking (TDT) to temporal analysis and new techniques are steadily being incorporated into the field of history. Temporal analysis with machine learning algorithms have been able to organize documents into a hierarchical system based on time by using a technique called temporal tagging. Temporal tagging analyzes a document to establish a timeline for an event by detecting relative, implicit, and explicit temporal descriptions.

An additional benefit of computational history is that it can be used for checking the credibility of a historical source since historical narratives are often crafted to fit the political and identity narratives of those in power. Harnessing the advantages of machine learning algorithms, digging through large sets of historical documents can assist in a thorough analysis of historical evidence by finding other pieces of documents that can corroborate the claims of the original source.

Text Mining

Text mining is the process of deriving information from text. It involves the extraction of information from sources such as books, articles, and websites. The earliest use of text mining in the field of humanities is often credited to Robert Busa, who conceived of and oversaw the Index Thomisticus project. Busa’s project digitized 179 texts of Thomas Aquinas over 34 years, which totaled over 10 million words.

With the proliferation of computers and computational techniques over the last few decades, humanities researchers have increasingly been using computational methods to assist them with textual analysis. In addition to simple methods like lexical analysis, modern humanities researchers have increasingly used a combination of statistical and machine learning techniques to carry out textual analyses. Scholars commonly employ methods such as text categorization, text clustering, and sentiment analysis.

Note: I would not be so extensive in this portion of the article since there is an entire separate Text Mining Wikipedia page

Natural Language Processing of Historical Texts

While many types of text mining, i.e. biomedical, are motivated by commercial applications, Natural Language Processing (NLP) of historical texts is often motivated by humanities-based research. The first step for computational analysis of historical documents is to digitize the journals and books held at museums and libraries and there are several partnerships engaging in such work. For example in 2004, Stanford, the University of Michigan, Harvard, Oxford, and the New York Public Library partnered with the Google Library Project in a digitization project.

After text has been translated into a digital format, it must then be translated into a format that is capable of being processed by a computer. One primary approach is to map characters of a script to a set of positive integers, which gives you an encoded character set, the standard encoded set being ASCII.

Modeling the Past

There are large amounts of untapped information in historical texts (note: include examples?), and one way to make use of them is by a technique known as "predicting the past". (Reference?) Researchers can better evaluate current models by overlaying current analytics with historical information derived from natural language processing of historical documents.

While drawing retrodictions or predictions about unexplored parts of history is a natural consequence of exploring the past, this process remains controversial. Using statistics to explain human behavior is said to not fully provide an explanation or a level of understanding often considered necessary in research disciplines.

Temporal Information Retrieval

Recent scholars have begun to use text analysis methods to exploit the temporal information present in historical documents. Such scholars use natural language processing to extract temporal expressions from documents, on which they perform exploratory searches, clustering, similarity studies, and other analyses. Temporal techniques give scholars the ability to identify temporal-based trends in the documents they study. (Reference?)

Machine Learning Applications

In recent years, machine learning methods have gained greater prevalence in the field of history. One area where machine learning has been applied is in artifact reconstruction. By training models to learn the process of degradation, for example, scholars have used AI models to reassemble eroded artifacts and damaged documents.

Edit from Jiatong Li (Logen)
One such example involves utilizing text mining to analyze how history has been recorded in the past. Advancements in computational techniques have allowed historians to parse through massive amounts of data while minimizing the manpower required and have opened up new fields of studies like computational linguistics and computational social science [Consider adding Computational social science link]. As computational history involves utilizing emerging computational techniques to view history, creative and new approaches are constantly created. A study by a team at Hokkaido University created a new way of distinguishing the historical importance of a figure by utilizing an algorithm similar to PageRank that examined Wikipedia links found on the page of a historical figure. Other computational techniques that are related to the field of computational history include anywhere from topic detection and tracking (TDT) to temporal analysis and new techniques are steadily being incorporated into the field of history. Temporal analysis with machine learning algorithms have been able to organize documents into a hierarchical system based on time by using a technique called temporal tagging. Temporal tagging analyzes a document to establish a timeline for an event by detecting relative, implicit, and explicit temporal descriptions.

''An additional benefit of computational history is that it can be used for checking the credibility of a historical source since historical narratives are often crafted to fit the political and identity narratives of those in power. Harnessing the advantages of machine learning algorithms, digging through large sets of historical documents can assist in a thorough analysis of historical evidence by finding other pieces of documents that can corroborate the claims of the original source. [Is this paragraph relevant to the subsections you listed below? I suggest either adding a subsection on the benefit of computational history or drop it'']

Text Mining

Text mining is the process of deriving information from text. It involves the extraction of information from sources such as books, articles, and websites. The earliest use of text mining in the field of humanities is often credited to Robert Busa [Consider adding Roberto Busa wiki link], who conceived of and oversaw the Index Thomisticus project. Busa’s project digitized 179 texts of Thomas Aquinas over 34 years, which totaled over 10 million words.

With the proliferation of computers and computational techniques over the last few decades, humanities researchers have increasingly been using computational methods to assist them with textual analysis. In addition to simple methods like lexical analysis, modern humanities researchers have increasingly used a combination of statistical and machine learning techniques to carry out textual analyses. Scholars commonly employ methods such as text categorization, text clustering, and sentiment analysis.

Natural Language Processing of Historical Texts

While many types of text mining, i.e. biomedical, are motivated by commercial applications [Any reference to this argument?], Natural Language Processing of historical texts is often motivated by humanities-based research. The first step for computational analysis of historical documents is to digitize the journals and books held at museums and libraries and there are several partnerships engaging in such work. For example in 2004, Stanford, the University of Michigan, Harvard, Oxford, and the New York Public Library partnered with the Google Library Project in a digitization project. [I suggest to keep this sentence simple since digitization is not related to NLP"

After text has been translated into a digital format, it must then be translated into a format that is capable of being processed by a computer. One primary approach is to map characters of a script to a set of positive integers, which gives you an encoded character set, the standard encoded set being ASCII.

Modeling the Past

There are large amounts of untapped information in historical texts, and one way to make use of them is by a technique known as "predicting the past". Researchers can better evaluate current models by overlaying current analytics with historical information derived from natural language processing of historical documents.

While drawing retrodictions or predictions about unexplored parts of history is a natural consequence of exploring the past, this process remains controversial. Using statistics to explain human behavior is said to not fully provide an explanation or a level of understanding often considered necessary in research disciplines.

Temporal Information Retrieval

Recent scholars have begun to use text analysis methods to exploit the temporal information present in historical documents. Such scholars use natural language processing to extract temporal expressions from documents, on which they perform exploratory searches, clustering, similarity studies, and other analyses. Temporal techniques give scholars the ability to identify temporal-based trends in the documents they study.

[Random thought: I think you can check spatial data analysis: Spatial analysis. That will be very helpful in explaining what temporal information retrieval means.]

Machine Learning Applications

In recent years, machine learning methods have gained greater prevalence in the field of history. One area where machine learning has been applied is in artifact reconstruction. By training models to learn the process of degradation, for example, scholars have used AI models to reassemble eroded artifacts and damaged documents. [Like this reference. Very updated!]

Peer Review (Steph Ran)

One such example involves utilizing text mining to analyze how history has been recorded in the past. Advancements in computational techniques have allowed historians to parse through massive amounts of data while minimizing the manpower required and have opened up new fields of studies like computational linguistics and computational social science. '''[A minor comment on sentence structure. Maybe trying something like “Advancements in computational techniques have opened up new fields such as computational linguistics and computational social science while minimizing the necessary manpower and allowing historians to parse through massive amounts of data.” could make it slightly more clear and concise!]''' As computational history involves utilizing emerging computational techniques to view history, creative and new approaches are constantly created. A study by a team at Hokkaido University created a new way of distinguishing the historical importance of a figure by utilizing an algorithm similar to PageRank that examined Wikipedia links found on the page of a historical figure. Other computational techniques that are related to the field of computational history include anywhere [Not sure if “anywhere” is the right word to use here-- maybe try “methods”?] from topic detection and tracking (TDT) to temporal analysis and new techniques are steadily being incorporated into the field of history. Temporal analysis with machine learning algorithms have been able to organize documents into a hierarchical system based on time by using a technique called temporal tagging. Temporal tagging analyzes a document to establish a timeline for an event by detecting relative, implicit, and explicit temporal descriptions.

An additional benefit '''[This could be interpreted as bias, so it could help to keep things objective here. Maybe something like “Another use of computational history is for checking the credibility…”.]''' of computational history is that it can be used for checking the credibility of a historical source since historical narratives are often crafted to fit the political and identity narratives of those in power. Harnessing the advantages of machine learning algorithms, digging through large sets of historical documents can assist in a thorough analysis of historical evidence by finding other pieces of documents that can corroborate the claims of the original source.

Text Mining

Text mining is the process of deriving information from text. It involves the extraction of information from sources such as books, articles, and websites. The earliest use of text mining in the field of humanities is often credited to Robert Busa [Is there a hyperlink that can be included here for Robert Busa?], who conceived of and oversaw the Index Thomisticus project [And here!]. Busa’s project digitized 179 texts of Thomas Aquinas over 34 years, which totaled over 10 million words.

With the proliferation of computers and computational techniques over the last few decades, humanities researchers have increasingly been using computational methods to assist them with textual analysis. In addition to simple methods like lexical analysis, modern humanities researchers have increasingly used a combination of statistical and machine learning techniques to carry out textual analyses. Scholars commonly employ methods such as text categorization, text clustering, and sentiment analysis.

Natural Language Processing of Historical Texts

While many types of text mining, i.e. biomedical, are motivated by commercial applications, Natural Language Processing of historical texts is often motivated by humanities-based research. The first step for computational analysis of historical documents is to digitize the journals and books held at museums and libraries and there are several partnerships engaging in such work. For example in 2004, Stanford, the University of Michigan, Harvard, Oxford, and the New York Public Library partnered with the Google Library Project in a digitization project.

After text has been translated into a digital format, it must then be translated into a format that is capable of being processed by a computer. One primary approach is to map characters of a script to a set of positive integers, which gives you an encoded character set, the standard encoded set being ASCII.

Modeling the Past

There are large amounts of untapped information in historical texts, and one way to make use of them is by a technique known as "predicting the past". Researchers can better evaluate current models by overlaying current analytics with historical information derived from natural language processing of historical documents.

While drawing retrodictions or predictions about unexplored parts of history is a natural consequence of exploring the past, this process remains controversial. Using statistics to explain human behavior is said to not fully provide an explanation or a level of understanding often considered necessary in research disciplines.

Temporal Information Retrieval

Recent scholars have begun to use text analysis methods to exploit the temporal information present in historical documents. Such scholars use natural language processing to extract temporal expressions from documents, on which they perform exploratory searches, clustering, similarity studies, and other analyses. Temporal techniques give scholars the ability to identify temporal-based trends in the documents they study.

Machine Learning Applications

In recent years, machine learning methods have gained greater prevalence in the field of history. One area where machine learning has been applied is in artifact reconstruction. By training models to learn the process of degradation, for example, scholars have used AI models to reassemble eroded artifacts and damaged documents.

'''[This was a fantastic overview! As someone with little background in this topic, I understood your explanation as it was very thorough and clear. I do wonder if it would be more beneficial to go in depth to one or two of the subtopics you address? There were quite a few, which is awesome, but it could potentially be more focused if fewer topics were addressed in more depth.]'''

Edits by Jackie Caraveo
Computational History -- Jack, James, Yamato

Computational History (not to be confused with computation history), sometimes also called Histoinformatics, is a multidisciplinary field that studies history through machine learning and other data-driven, computational approaches. (UP TO HERE IS THE ORIGINAL PAGE)

(NEW CONTENT)

One such example involves utilizing text mining to analyze how history has been recorded in the past. Advancements in computational techniques have allowed historians to parse through massive amounts of data while minimizing the manpower required and have opened up new fields of studies like computational linguistics and computational social science. As computational history involves utilizing emerging computational techniques to view history, creative and new approaches are constantly created. A study by a team at Hokkaido University created a new way of distinguishing the historical importance of a figure by utilizing an algorithm similar to PageRank that examined Wikipedia links found on the page of a historical figure.[''Make sure to cite the study at the end of the sentence. Good job paraphrasing!]'' Other computational techniques that are related to the field of computational history include anywhere from topic detection and tracking (TDT) to temporal analysis and new techniques are steadily being incorporated into the field of history. Temporal analysis with machine learning algorithms have been able to organize documents into a hierarchical system based on time by using a technique called temporal tagging. Temporal tagging analyzes a document to establish a timeline for an event by detecting relative, implicit, and explicit temporal descriptions.

An additional benefit [avoid using terms like "benefit" because it can be interpreted as an opinion] of computational history is that it can be used for checking the credibility of a historical source since historical narratives are often crafted to fit the political and identity narratives of those in power. Harnessing the advantages of machine learning algorithms, digging through large sets of historical documents can assist in a thorough analysis of historical evidence by finding other pieces of documents that can corroborate the claims of the original source.

Text Mining

Text mining is the process of deriving information from text. It involves the extraction of information from sources such as books, articles, and websites. The earliest use of text mining in the field of humanities is often credited to Robert Busa, who conceived of and oversaw the Index Thomisticus project. Busa’s project digitized 179 texts of Thomas Aquinas over 34 years, which totaled over 10 million words. [I think it'd be interesting to hyperlink Robert Busa and his work here]

With the proliferation of computers and computational techniques over the last few decades, humanities researchers have increasingly been using computational methods to assist them with textual analysis. In addition to simple methods like lexical analysis, modern humanities researchers have increasingly used a combination of statistical and machine learning techniques to carry out textual analyses. Scholars commonly employ methods such as text categorization, text clustering, and sentiment analysis.

Natural Language Processing of Historical Texts

While many types of text mining, i.e. biomedical, are motivated by commercial applications, Natural Language Processing of historical texts is often motivated by humanities-based research. The first step for computational analysis of historical documents is to digitize the journals and books held at museums and libraries and there are several partnerships engaging in such work. For example in 2004, Stanford, the University of Michigan, Harvard, Oxford, and the New York Public Library partnered with the Google Library Project in a digitization project. [add link here to reference this partnership]

After text has been translated into a digital format, it must then be translated into a format that is capable of being processed by a computer. One primary approach is to map characters of a script to a set of positive integers, which gives you an encoded character set, the standard encoded set being ASCII.

Modeling the Past

There are large amounts of untapped information in historical texts, and one way to make use of them is by a technique known as "predicting the past". Researchers can better evaluate current models by overlaying current analytics with historical information derived from natural language processing of historical documents.

While drawing retrodictions or predictions about unexplored parts of history is a natural consequence of exploring the past, this process remains controversial. Using statistics to explain human behavior is said to not fully provide an explanation or a level of understanding often considered necessary in research disciplines.

Temporal Information Retrieval

Recent scholars have begun to use text analysis methods to exploit the temporal information present in historical documents. Such scholars use natural language processing to extract temporal expressions from documents, on which they perform exploratory searches, clustering, similarity studies, and other analyses. Temporal techniques give scholars the ability to identify temporal-based trends in the documents they study.

Machine Learning Applications

In recent years, machine learning methods have gained greater prevalence in the field of history. One area where machine learning has been applied is in artifact reconstruction. By training models to learn the process of degradation, for example, scholars have used AI models to reassemble eroded artifacts and damaged documents.

Notes: This is great! I'd suggest adding a bit more detail to the last two mini topics or maybe getting rid of them to add more detail to the other topics.

Peer Review by George Afentakis
Computational History (not to be confused with computation history), sometimes also called Histoinformatics, is a multidisciplinary field that studies history through machine learning and other data-driven, computational approaches. (UP TO HERE IS THE ORIGINAL PAGE)

(NEW CONTENT)

One such example involves utilizing text mining to analyze how history has been recorded in the past. Advancements in computational techniques have allowed historians to parse through massive amounts of data while minimizing the manpower required and have opened up new fields of studies like computational linguistics and computational social science. As computational history involves utilizing emerging computational techniques to view history, creative and new approaches are constantly created. A study by a team at Hokkaido University created a new way of distinguishing the historical importance of a figure by utilizing an algorithm similar to PageRank that examined Wikipedia links found on the page of a historical figure. Other computational techniques that are related to the field of computational history include anywhere from topic detection and tracking (TDT) to temporal analysis and new techniques are steadily being incorporated into the field of history. Temporal analysis with machine learning algorithms have been able to organize documents into a hierarchical system based on time by using a technique called temporal tagging. Temporal tagging analyzes a document to establish a timeline for an event by detecting relative, implicit, and explicit temporal descriptions.

An additional benefit of computational history is that it can be used for checking the credibility of a historical source since historical narratives are often crafted to fit the political and identity narratives of those in power. Harnessing the advantages of machine learning algorithms, digging through large sets of historical documents can assist in a thorough analysis of historical evidence by finding other pieces of documents that can corroborate the claims of the original source.

Text Mining

Text mining is the process of deriving information from text. It involves the extraction of information from sources such as books, articles, and websites. The earliest use of text mining in the field of humanities is often credited to Robert Busa, who conceived of and oversaw the Index Thomisticus project. Busa’s project digitized 179 texts of Thomas Aquinas over 34 years, which totaled over 10 million words. [I think Robert Busa has a wikipedia page so it might be helpful to add him as a hyperlink and I think the same goes for for Index Thomisticus (also I think his name might be Roberto)]

With the proliferation of computers and computational techniques over the last few decades, humanities researchers have increasingly been using computational methods to assist them with textual analysis. In addition to simple methods like lexical analysis, modern humanities researchers have increasingly used a combination of statistical and machine learning techniques to carry out textual analyses. Scholars commonly employ methods such as text categorization, text clustering, and sentiment analysis.

'''[Maybe you should incorporate the information in the text mining section in the first paragraph where you first mention text mining. I think this would make for an easier read.]'''

Natural Language Processing of Historical Texts

While many types of text mining, i.e. biomedical, are motivated by commercial applications, Natural Language Processing of historical texts is often motivated by humanities-based research. The first step for computational analysis of historical documents is to digitize the journals and books held at museums and libraries and there are several partnerships engaging in such work. For example in 2004, Stanford, the University of Michigan, Harvard, Oxford, and the New York Public Library partnered with the Google Library Project in a digitization project.

After text has been translated into a digital format, it must then be translated into a format that is capable of being processed by a computer. One primary approach is to map characters of a script to a set of positive integers, which gives you an encoded character set, the standard encoded set being ASCII.

Modeling the Past

There are large amounts of untapped information in historical texts, and one way to make use of them is by a technique known as "predicting the past". Researchers can better evaluate current models by overlaying current analytics with historical information derived from natural language processing of historical documents.

While drawing retrodictions or predictions about unexplored parts of history is a natural consequence of exploring the past, this process remains controversial. Using statistics to explain human behavior is said to not fully provide an explanation or a level of understanding often considered necessary in research disciplines.

Temporal Information Retrieval

Recent scholars have begun to use text analysis methods to exploit the temporal information present in historical documents. Such scholars use natural language processing to extract temporal expressions from documents, on which they perform exploratory searches, clustering, similarity studies, and other analyses. Temporal techniques give scholars the ability to identify temporal-based trends in the documents they study.

Machine Learning Applications

In recent years, machine learning methods have gained greater prevalence in the field of history. One area where machine learning has been applied is in artifact reconstruction. By training models to learn the process of degradation, for example, scholars have used AI models to reassemble eroded artifacts and damaged documents.

'''[These last sections are great but they are a bit small. I would recommend either adding to them or somehow merging them so they are not separate sections. ]'''

'''[Overall Thoughts: Great Work! The only thing I might recommend changing is combining the last few smaller sections into one big section. Either that or you can add a bit more context to them and make them bigger]'''

Peer Review by Yuri Vieira Sugano
Computational History -- Jack, James, Yamato

Computational History (not to be confused with computation history), sometimes also called Histoinformatics, is a multidisciplinary field that studies history through machine learning and other data-driven, computational approaches. (UP TO HERE IS THE ORIGINAL PAGE)

(NEW CONTENT)

One such example involves utilizing text mining to analyze how history has been recorded in the past. (being naive to the topic, I am not sure if I understand the example. From the article it sounds like the authors are not just trying to understand how history was recorded, but rather use how history has been recorded to gain insight about memories or societal views. Perhaps as an opening example, you can further explain what the technology can actually accomplish) Advancements in computational techniques have allowed historians to parse through massive amounts of data while minimizing the manpower required and have opened up new fields of studies like computational linguistics and computational social science. As computational history involves utilizing emerging computational techniques to view history, creative and new approaches are constantly created. A study by a team at Hokkaido University created a new way of distinguishing the historical importance of a figure by utilizing an algorithm similar to PageRank that examined Wikipedia links found on the page of a historical figure. (I think the study is not necessary relevant to the introduction, but its findings are. Perhaps rephrase to: "Computational history can provide insight about the historical importance of a figure[2]." Other computational techniques that are related to the field of computational history include anywhere from topic detection and tracking (TDT) to temporal analysis and new techniques are steadily being incorporated into the field of history. (perhaps be a bit more specific, are all these techniques computational, or specific to machine learning?) Temporal analysis with machine learning algorithms have been able to organize documents into a hierarchical system based on time by using a technique called temporal tagging. Temporal tagging analyzes a document to establish a timeline for an event by detecting relative, implicit, and explicit temporal descriptions.

An additional benefit of computational history (Rephrase to "Another use of computational technology" - you have not introduced one benefit to introduce an additional, and I agree with Steph that benefit carries personal opinion) is that it can be used for checking the credibility of a historical source since historical narratives are often crafted to fit the political and identity narratives of those in power. (Again, I think an additional sentence describing how historical narratives inform credibility of sources would be helpful). Harnessing the advantages of machine learning algorithms, digging through large sets of historical documents can assist in a thorough analysis of historical evidence by finding other pieces of documents that can corroborate the claims of the original source.

(Perhaps rephrase paragraph to "Analysis of large sets of historical documents can assist in a thorough analysis of historical evidence. Individual historical sources can be compared to each other and their credibility will emerge from these comparisons. )

Text Mining

Text mining is the process of deriving information from text. It involves the extraction of information from sources such as books, articles, and websites (suggestion: "books, articles, and websites, which are often the source of historical documents"). The earliest use of text mining in the field of humanities is often credited to Robert Busa, who conceived of and oversaw the Index Thomisticus project. Busa’s project digitized 179 texts of Thomas Aquinas over 34 years, which totaled over 10 million words.

With the proliferation of computers and computational techniques over the last few decades, humanities researchers have increasingly been using computational methods to assist them with textual analysis. (Rephrase suggestion: "Researchers have been increasingly assisted by computational methods, which have proliferated over the last few decades" ) In addition to simple methods like lexical analysis, modern humanities researchers have increasingly used a combination of statistical and machine learning techniques to carry out textual analyses. Scholars commonly employ methods such as text categorization, text clustering, and sentiment analysis. (Check whether these concepts have their own unique Wiki pages -- sometimes it is helpful to hover over a cross-reference to get a small idea of what it is about)

Natural Language Processing of Historical Texts

While many types of text mining, i.e. biomedical, are motivated by commercial applications, Natural Language Processing of historical texts is often motivated by humanities-based research. The first step for computational analysis of historical documents is to digitize the journals and books held at museums and libraries and there are several partnerships engaging in such work. For example in 2004, Stanford, the University of Michigan, Harvard, Oxford, and the New York Public Library partnered with the Google Library Project in a digitization project.

After text has been translated into a digital format, it must then be translated into a format that is capable of being processed by a computer. One primary approach is to map characters of a script to a set of positive integers, which gives you an encoded character set, the standard encoded set being ASCII.

Modeling the Past

There are large amounts of untapped information in historical texts, and one way to make use of them is by a technique known as "predicting the past". Researchers can better evaluate current models by overlaying current analytics with historical information derived from natural language processing of historical documents.

While drawing retrodictions or predictions about unexplored parts of history is a natural consequence of exploring the past, this process remains controversial. Using statistics to explain human behavior is said to not fully provide an explanation or a level of understanding often considered necessary in research disciplines.

Temporal Information Retrieval

Recent scholars have begun to use text analysis methods to exploit the temporal information present in historical documents. Such scholars use natural language processing to extract temporal expressions from documents, on which they perform exploratory searches, clustering, similarity studies, and other analyses. Temporal techniques give scholars the ability to identify temporal-based trends in the documents they study.

Machine Learning Applications

In recent years, machine learning methods have gained greater prevalence in the field of history. One area where machine learning has been applied is in artifact reconstruction. By training models to learn the process of degradation, for example, scholars have used AI models to reassemble eroded artifacts and damaged documents.

(I think this is a great summary, I had never heard of computational history and the read was informative, interesting and engaging. The only overall suggestion is to perhaps think a bit about the subcategories. Right now, they are not informative. One suggestion might be to group them into "Techniques" (under which you will have text mining and natural language processing and machine learning) and "Applications" (under which you will have modeling the past, information retrieval and reassembly of eroded artifacts and damaged documents).