Development of a complete essay writing self study system for junior high school - including systems for Chinese essay writing Scaffolding Instruction, real-time writing error analysis, essay assessment with correction, and writing improvement.

Abstract

Currently, there is no essay correction system for Chinese native speakers in the market. Most of today's essay correction systems are still at the experimental stage. Besides they are designed for teachers assessing essays. As a result, students only know their score result but they do not know how to improve their writing skill. It is limited to help the students due to the lack of essay writing assisting system.

This project develops the first writing system, which especially designed to fulfill students needs. This project not only considers teacher's needs of the error detection and assessment also offers counseling and remedial systems for writing shortcomings. Technological innovation and how a self learning system can be developed according to the needs of business partners are detailed below:

1 Content Innovation: This project will design new features following the writing process of students, from the "essay preparation phase" , "the essay writing process" to the  "writing completion".

(1) Pre-writing activities:

(1a) Having students mark essays of their peers, and then letting them compare their marking to those of a teachers on the same essay will allow the students to learn from the mistakes made by others and so improve their own writing (Refer to project plan for details: 12.2.1).

(1b) The System provides students with instruction scaffolding while writing, in order to help students organize and structure their writing (Refer to project plan for details: 12.2.1).

(2) Essay writing stage: While composing their essay, student can use the following functions:

(2a) Using "the real-time writing shortcomings detection and prompting system" to identify, provide propomtly the required sentences for the essay writing, and detect in real-time the mistakes been made (refer to project plan for details: 12.2.2).

(2b) Providing the students with the necessary sentences, which have been corrected by teachers..

(3) Essay writing late stage (after auto-marking, in which the system offers a remedial mode to correct errors)

(3a) The system provides language use explanation through context or by means of narrative movies: when the expressions used by students are used in the wrong context, the system will be connected with the digital teaching materisl of our coorperation  partners’s "Taiwan Knowledge Bank" which will provide "teaching videos explaining" for the use of the expression, it will also connect to youtube, so that students understand the meaning and the context in which the term is used.

(3b) Using game task-based teaching model and adaptive tests to determine whether students have the ability to correct their mistakes: At first, the task-based games test whether students already have the necessary knowledge to correct their own mistakes. And then re-test using various level exercises to establish whether students have acquired and solidified using the films, dictionaries and learning materials the required knowledge (see project plan for details 12.4).

2.     Technological innovation:

(1)     The first system to achieve automatic essay correction : The system makes use of a "Teacher essay correction Knowledge base (a teacher correction model database)", and Machine Learning algorithm to achieve automatic essay correction in order to reduce the teachers ' correcting essay workload.

(2)     The System is Taiwan's first emulation of a real person marking and assessing an essay. For example, it uses the comprehensive assessment program for junior high school students, which are classified as four categories according to students' essays ability in regard of approach and material , structure and organization, wording and phrasing, "character errors, format, and punctuation"), also rhetorical skills. Conjunctions and complex sentences are keypoints to render logical organization of the essay and have an impact on the essay assessment. Therefore, they are part of the assessment knowledge base. Natural Language Processing technology is used to implement diversified assessment module, thus enabling the proposed system to be more complete than existing essay assessment systems (refer to project plan for details : 12.2.3.2 to 12.2.3.7).

These four essay assessment categories are not random requirements. They are required elements used in the process of writing an essay. Therefore, these assessment categories cannot be easily changed.

(3) Current essay error detection only deals with words that are close together. and it can’t process words that are not immediately adjacent or errors that are beyond the scope of a sentence (for example compound sentences). The proposed system is the first try to integrate shallow sentence parsing and Language Model to deal with those errors that are not processed by current systems.

(4) Built-in data auto-update mechanisms allow the users, while using the system i.e. writing essays, at the same time to help the system updateing its content and its knowledge base in order to effectively improve operation(refer to project plan for details :12.6)。

 (5) The project also aims to grasp crucial errors when teachers assess essays. Some so-called critical error are

(a)    Errors that affect the teacher’s assessment most,

(b)    Errors repeating in a higher frequency,

(c)    Errors that are not handled by current Chinese essay systems (refer to details at  12.2.3.7).

3. Cooperating with the company "Taiwan Knowledge Bank":

The development of this innovative essay writing system, enabling the company to gain a competitive edge. Currently in the domain of mathematic, the company has offered a "creative robot Course". In the language domain, they has also been planning a "self learning essay writing system". However. there is no language professional taking part in its development. Combining this projects self-learning essay writing system with the business partners original teaching modules, we can achieve a complete professional language teaching system: The business partners language teaching materials mainly aims at teaching and learning, lacking of  assessment of the students' writing ability and the ability to correct students mistakes. Therefore, the proposed system can accordingly combine with the business partners teaching materials in the early stage. Also. at a later stage of the writing process the proposed system can offer errors marking, assessment and remedial measures, completing the business partner systems deficiencies in order to obtain a complete professional language teaching system.

 The project, "Complete essay writing self study system for junior high school", includes the following systems: "essay writing teaching and tutoring platform", "essay writing scaffolding instruction", "System for real-time detection of errors and correction", "system for automatic detection of error in the essay","Automatic essay grading system" ,"Automatic essay  correcting system"  and "essay tutoring and writing improvement system."

 During the project execution period, we will use the project management milestone method. At first, we will set key output product including (a)the analysis of the requirements for completion of the each subsystems and modules, (b) systems analysis, (c) system design, (d) and system implementation. Besides, at each stage the teams involved in the project will be audited for their achievement. If the progresses fall behind schedule we’ll initiate the necessary corrective measures.

 Keywords: Essay writing self-learning system, Essay automated assessment, Essay writing tutoring system, Essay writing correction system, essay writing marking system, Essay writing error detection system, Essay writing correction database, Essay corpus.